A data-driven approach to measure money laundering risk and its relationship with corruption

Abstract

The issue of how the money laundering (ML) risk of countries is measured is not a merely technical problem. It has a strong impact on countries’ development, in particular of the Global South. First, because official anti-money laundering (AML) blacklists and grey lists, which frequently include developing and small countries, are heavily questioned in terms of their capacity to represent actual ML risk. Second, because a proper measurement of ML risk would allow to better appreciate the relationship between ML and other crimes, in particular corruption, which represent key threats for most of the Global South and the developing world. This paper addresses this issue by proposing a methodology which, based on a data-driven approach, can be employed to assess the actual risk of ML at country level, and to analyse its relationship with transnational corruption. In particular it proposes a composite indicator of ML risk based on the inputs from previous economic, sociological and criminological literature. These refer to key threats and vulnerabilities and are operationalised into one or more measurable proxy variables. The paper then validates the indicator by applying it to selected countries, and by comparing it with (a) observed evidence of ML and (b) measures of corruption and corruption exposure. Results show a strong correlation between the new risk indicator and observed evidence of ML, but no correlation with corruption levels. However, countries at high risk of ML appear to be more exposed to investors and beneficial owners from countries at higher level of corruption, suggesting they can also attract corruption money. Negative correlation with official AML blacklists is observed: surprisingly, listed countries show lower ML risk on average. The work may help to revisit the current AML blacklisting process, and minimise its unintended consequences such as de-risking on smaller countries and Global South as a whole.