Improving the resilience of systems
Systemic risk describes the likelihood of cascading failures in networks. Such risks arise in a broad range of different systems, such as power grids, ecosystems, supply chains, financial networks, disease dynamics, and transportation networks. The Systemic Risk and Network Dynamics (SRND) cross-cutting project at IIASA, aims to develop capabilities for analyzing systemic risks and to demonstrate how to assess and mitigate risks of cascading failures.
While most existing approaches to systemic-risk assessment are application-specific, similarities between systems offer great potential for cross-fertilization and synergetic analyses. Specifically, the project is developing cross-cutting measures of systemic risk, prognostic tools for assessing the likelihood and extent of cascading collapses under uncertainty, methods for reducing systemic risk through network design and control, and new approaches to the governance of systemic risk.
The project explores systemic risk in a broad range of applications from natural to human-made systems. In 2017 for example, researchers working on the SRND project explored the effectiveness of credit default swaps (CDSs) as an alternative or complementary instrument to the systemic risk tax studied earlier . Over recent years, CDSs have acquired a negative reputation, as they are widely used for speculations, which are seen as exacerbating financial systemic risk. However, using an economic-financial model, the results of one study  showed that, by properly shifting financial exposures from one institution to another, a CDS market can be designed to rewire the network of interbank exposures in ways that make it more resilient to insolvency cascades.
The project also developed and used an agent-based model (ABM) simulating a national economy previously developed by its researchers, to estimate the indirect economic consequences of direct losses arising from floods. This model is the first to use a 1:1 scale to represent a country’s natural persons and legal entities, such as firms and banks, and to simulate their interactions. The ABM is currently calibrated for Austria, using data from national accounts, census data, and business information. It is driven by a probabilistic flood model, which uses the copula approach to predict flood losses while accounting for spatial dependencies. In this way, the researchers link environmental and economic processes in a nationwide simulation. Their analysis predicts that moderate floods induce positive indirect economic effects in the short and medium term, and small but negative indirect economic effects in the long term. In contrast, large-scale floods result in a more pronounced negative economic response in the long term. This approach allows the researchers to identify winners and losers in unprecedented detail across all economic sectors, as well as fiscal consequences for the government, both of which are crucial for managing extreme events resulting from climate change and natural disasters. A paper presenting these findings has been submitted for publication.
Furthermore, the project’s work on systemic risks in ecosystems is ongoing. Researchers working in this thematic area analyze how species losses propagate through food webs. In particular, they have developed what has become the world’s largest database of quantified food webs, and have used this information on ecosystems from around the globe to calibrate their models. This provides a unique basis for addressing controversies that have persisted in the ecological community for decades, concerning the question of which structural features make food webs more or less vulnerable to species loss. The researchers have expanded this resilience analysis from the ecosystem level to the species level.
Since its inception, the SRND project has been enabling the three participating IIASA programs to pool their methodological expertise on dynamic systems, risk analysis, and network theory. In light of this, a perspective paper is being prepared that presents an integrated approach using the copula methodology, for combining individual risks (in the form of probabilistic distributions) and systemic risks (in the form of copulas describing the dependencies among such distributions). This approach is especially useful when extreme events (occurring at low probabilities, but having high impacts) that affect agents in a system can lead to a tightening of the connections between some or all agents, as is often the case in, for example, financial systemic risks.
 Poledna S, Bochmann O, & Thurner S (2017). Basel III capital surcharges for G-SIBs are far less effective in managing systemic risk in comparison to network-based, systemic risk-dependent financial transaction taxes. Journal of Economic Dynamics and Control 77: 230–246.
 Leduc MV, Poledna S, & Thurner S (2017). Systemic risk management in financial networks with credit default swaps. Journal of Network Theory in Finance 3 (3): 19–39.
Top image © Vintage Tone | Shutterstock