Designing early warning systems for detecting systemic risk: A case study and discussion

Systemic risks are potential trigger events or developments that could undermine the viability of entire networks or systems. Growing complexity in systems make such risks both more likely to occur and more difficult to anticipate. The tools for detecting systemic risk have not kept pace with these challenges; traditional methods are too intermittent, too slow, and too narrow in focus for timely systemic risk detection. However, recent developments in big data analysis and artificial intelligence (AI) have the potential to revolutionize Early Warning Systems (EWSs) for detecting systemic risk. EWSs that are supported by these technologies could provide users with earlier warning signals of a wider range of risks and more up-to-date measures of the fragility of the system against these risks. This area of research is nascent and lacks a robust methodology for designing such EWSs. Addressing this issue, the present paper: 1) identifies the characteristics of competent EWSs; 2) outlines an approach for designing such EWSs; and 3) illustrates the value of this approach, by discussing the conceptual design of an EWS for detecting biosecurity incursions in the New Zealand pastoral industries.