What Counts as Corruption? Alternative Indicators for SDG16.5 in the Era of AI

The inclusionof anticorruption objectives among the Sustainable Development Goals (SDGs) must be celebrated. As then-Secretary-General Kofi Annan noted to the General Assembly in 1998, good governance could be “the single most important factor in eradicating poverty and promoting development”. Target 16.5, “Substantially reduce corruption and bribery in all their forms” materialises this objective within the SDGs, setting an vision of a (substantially) less corrupt world by 2030. However, as explored in this chapter, the quantification aspect of this target and its indicators, risk reinforcing stereotypes that see corruption as a global South (or “majority world”) issue and primes policymakers and other stakeholders to see all corruption through the bribery prism. The measurements to assess success in substantially reducing corruption and bribery in all their forms are two indicators tracking the proportion of persons (16.5.1) or businesses (16.5.2) who have paid (or were asked for) a bribe to a public official during the previous 12 months (Statistical Commission 2017, 30/71). This chapter focuses concretely on these indicators and contends that they are not adequate to assess changes in the incidence of corruption across countries. First, they use bribery as a proxy for corruption, introducing a metonymic approach that is not representative of the corruption reality in many countries. Second, the approach quantifies bribery based exclusively on the persons’ and firms’ direct experiences in their relations with public officials, introducing severe methodological limitations. Third, it builds into the indicators an intrinsic bias against global South countries. Bribes are unequivocally a much more common instance of corruption in everyday life in countries with lower per capita incomes. If we looked for corruption beyond bribery, the picture could be different. For example, there is some evidence that the incidence of favouritism may not be that dissimilar in developing and developed countries. Measuring those instances would better complement the bribery indicators. This chapter encourages us to consider a different quantification system based on big data analysis, which could be fairer, and easier to assess over the 15-year span of the SDGs. This could be possible with the use of Artificial Intelligence (AI). AI can be used to track the evolution of countries over time in a comparable manner. This makes the creation of a third indicator advisable and feasible. The chapter presents two possible candidates for that third indicator around procurement data and health-related payments. These measurements could be the basis of other complementary indicators in the final years of the SDGs or, more likely, a springboard towards a more responsive measurement of anticorruption progress post-2030.