Cleaning Up Our Financial System: Combating Money Laundering Using Multiparty Computation

Abstract

Globally, it is estimated by the UN that 2% - 5% of the annual GDP is lost to money laundering. Cur-rent anti-money laundering efforts are hindered by both the lack of trust between financial institutions internationally and the presence of local privacy regulations like GDPR. This makes it unfeasible to share plain transaction data between financial institutions internationally. Secure multiparty computation is a cryptographic technique that enables a set of parties to interact and compute a joint function of their private inputs while revealing nothing but the output. Thus, MPC has the potential to facilitate greater collaboration between financial institutions and governmental organisations internationally and upscale anti-money laundering efforts securely. In this paper we aim to explore how MPC could be used to improve current anti-money laundering detection techniques. This is done by providing an overview of existing work in the field and proposing a new architecture that could be used to flag suspicious transaction. This architecture presents accounts and transactions as a social network and uses betweenness centrality to identify high-risk accounts. We outline how existing protocols can be used to build such a model and what further properties are to be considered to build even more sophisticated protocols.