Issue 01 / Foreword

Supersystem Risk and the End of the Anthropocene

James R. Watson, Laura E. R. Peters and Jamon Van Den Hoek


As we continue through the Anthropocene, we are witnessing ever-greater connectivity within and among our “world-systems”: financial markets, housing markets, social networks, transport systems, and ecosystems — all are becoming more and more connected through social, economic, and institutional links1,2,3. This connectivity brings with it greater economic efficiency (i.e. goods are produced at lower cost4, better availability of goods and services, as well as more potential for cooperation through larger social networks5. However, as many have noted, “connectivity is risk,” and the increased connectivity in and among our world-systems has led to a greater potential for global-scale catastrophe and reductions in human wellbeing6,7,8.

The connectedness of world-systems means that the nominal boundaries that we use to define a given world-system do not actually exist. For example, we think of the financial system as involving certain actors (e.g. stock and traders, companies, etc.), but many of these actors have strong (economic, social, institutional) ties to other actors that comprise different world-systems9. When world-systems themselves are intertwined, there exists a “supersystem” that is essentially our global socio-environmental system. As a consequence, looking beyond systemic risk (where a perturbation is confined to a single world-system) we must address “supersystem risk”. The 2020 pandemic is a case in point: transport systems enabled the rapid spread of the Covid-19 virus globally, the subsequent social distancing dismantled service industries, which due to the connectivity of financial markets, will likely lead to a global recession. The 2020 pandemic is creating numerous such knock-on events10,11, another being changes in the social network and mental health of individuals as social distancing continues to be in effect12. Ultimately, these cascading impacts of the pandemic will have uneven local consequences, with some people hit harder than others, for example through increasing prices of food and goods, and even food shortages as just-in-time food delivery systems break-down and food production diminishes due to labor shortages11.

Although vivid due to its recent and ongoing impacts, the 2020 pandemic is not alone as a global perturbation. The 2008 financial crisis occurred in part because of financial connections between the US housing and financial markets2,13; The Arab Spring revolution played out it in part due to prolonged regional drought and changes in grain prices14. These examples of supersystem risk being realized reveal aspects of our global socio-environmental system that must be recognized: 1) the socially produced risks we are facing are getting bigger, with larger impacts and greater reach, through increasing world-system connectivity — these emergent global-scale risks are termed supersystem risks 2) supersystem risks can be hidden: at any given instance they may not be structurally apparent, but as our world-systems change, they emerge; and perhaps most worryingly 3) these supersystem risks are being realized with increasing frequency, driven by an evolution of our world-systems toward ever-greater connectivity. Although the notion of the “Anthropocene” is still contested15,16, we now accept that human domination of our planet’s biosphere is in full swing. However, due to the economic incentives for increasing connectivity and the subsequent exposure to supersystem risks, perhaps our epoch will be short-lived. The production of hazards on massive scales is inherent to modern societies17, not the least of which is the supersystem itself, potentially placing a hard limit on the extent to which any society can develop.

Bigger risks: supersystem versus systemic risk

It is well recognized that the connectivity within and among our world-systems is increasing18. For example, in 2008, three months before the Lehman Brothers filed for bankruptcy, a paper was published describing the growth in connectivity of the banking system, being driven by interbank loans amongst other forms of connectivity, resulting in growing systemic risk19. A similar conversation has been ongoing with regards to the risk of a pandemic: many publications and talks20 have discussed in great depth how the globe was and continues to be (willfully) unprepared for a pandemic, and in particular the role of geographic connectivity driven by air-travel in promoting the likelihood of a pandemic21. All these discussions bring an emphasis to systemic risk7,22: that is the risk of not just one node failing (e.g. one bank becoming bankrupt, or one state suffering an epidemic) but the majority of all nodes failing (i.e. all banks failing, or all countries suffering a pandemic).

In today’s world it is almost impossible to think of a perturbation in one world-system being contained. Similarly, it is extremely difficult to predict where a perturbation that starts in one world-system may end up23. Not only does greater connectivity among world-systems mean that the impact of a perturbation has greater reach within and across systems9, it also means that the resilience (i.e. the ability to absorb and recover from a realized risk) of the overall supersystem is diminished and the likelihood of the entire supersystem suffering some sort of large structural change is increased. In general, three structural elements combine to determine the resilience of systems: diversity, modularity, and redundancy24,25. Each element has a unique contribution to system resilience, yet they cannot be optimized all at once. Indeed, through the need to be economically efficient, imitation and the desire for greater interpersonal and geographic connectivity, we have gained network connectivity but lost network modularity (i.e. the opposite of connectedness) and diversity (i.e. only a handful of key actors dominate most sectors3), and consequently our world-systems are now primed for large-scale transformative change. Transformative change is not necessarily bad – if a socio-environmental system is built to be overly extractive and/or destructive, then a large structural change could bring immediate as well as long-term benefits26. However, taken as a whole, the large shifts in the structural organization of our world-systems that result from a supersystem risk being realized will likely be unexpected and unplanned and as a consequence, they will likely engender abrupt drops in human wellbeing27. For example, recent work leveraging network theory28 has shown that greater connectivity in global trade networks has increased the potential for large and abrupt changes in the provision of food29,30. Unfortunately, it is evident that our world-systems are now so strongly connected that perturbations can spread far and have major impacts.

Hidden risks: complex adaptive system dynamics

The importance of the growing connectivity among our world-systems is mirrored by the dynamics occurring within them. Each world-system is a complex adaptive system, with strategic actors interacting over a range of spatial and temporal scales, and levels of organization. The dynamics of all kinds of complex adaptive systems, for example ecosystems, power-grids, social-networks...etc., emerge from the interactions between the actors that comprise the system (e.g. animals, power-busses, people, financial trading algorithms), and also from feedbacks with dynamics happening at higher levels of organization (e.g. herds, power-grids, communities, financial markets). This means a supersystem risk is realized through the actions of the actors comprising the system. This is in contrast to certain forms of “existential risk” that often involve a large external perturbation such as a planet-killer asteroid. Another related concept is that of femtorisks31: these are threats that result from the actions and interactions of actors that exist beneath the level of formal institutions. The term “femto” highlights the apparent insignificance of the individual actor that might be a source of such a risk. But, when embedded in a complex and adaptive system, the actions these small-scale actors might take can end up creating a cascade of events that have large-scale impacts, and in the case of supersystem risks, global impacts.

Another problem associated with the complex and adaptive nature of our world-systems is that supersystem risks can be hidden. Hidden risks are those that are not yet apparent, but that emerge as actors respond (often strategically) to the actions of one another. In the standard model of risk, one examines the probability of an event happening and its potential impact on the actor/system. The challenge with hidden risks is that at any given instant they may have a probability of zero. Thus, they are perceived to be inconsequential and more wickedly; they may not even be known yet. However, in complex adaptive systems, these probabilities are not constant. Instead they change over time as a function of the interaction of actors25,32. Thus, a hidden risk can emerge as the system evolves over time. The notion of hidden risks was captured by Donald Rumsfeld, the former U.S. Secretary of Defense, who described them as unknown unknowns, where we have no idea what the threat and risk actually are33. As our world-systems become more connected, the number of actors and links between them grow to such a point that it is almost impossible to identify when and how a supersystem risk might emerge. Those at the center of risk production are the only ones that can interpret the system well enough to propose solutions; but, where solutions are self-serving, they will not resolve the core of supersystem risk production.

More risks: the evolution of connectivity and the threat of collapse

Through increasing intra- and inter-connectedness, our world-systems have become more efficient locally, yet this has made them susceptible to supersystem risks. More importantly, they are continuing to evolve towards ever-greater connectivity, driven by the selection of actors and institutions that maximize economic efficiency. The problem is one of timescales. At relatively short timescales (e.g. from seconds to years) individuals, algorithms, businesses and governments are increasing the connectivity within and among our world-systems in order to increase economic efficiency. However, over longer timescales (e.g. decades and centuries), connectivity has increased to such a point that supersystem risks are now possible. Actors may realize that long-term viability requires resilience (i.e. through lower connectivity), but competition with their peers persuades them to focus their attention myopically and make decisions that maximize their near-term economic efficiency22,34. Indeed, in addition to growing connectivity, we have also witnessed a consolidation of wealth and influence among a small subset of actors that dominate any given world-system (i.e. transnational corporations3). This concentration of connectivity around a few major actors does not confer resilience. Quite the opposite: through a loss of diversity it makes the whole system more fragile.

Figure 1. Schematic showing the possible growth and eventual collapse of world-system connectivity. World-systems – social, transport, financial systems for example – are identified by the colored modules in the different networks. Initially these world-systems include connectivity within themselves, but, i) over time (i.e. decades, centuries) these world-systems have become highly connected, driven by incentives for economic efficiency. With greater connectivity comes supersystem risk and ii) at some point a super system risk is realized, and the connectivity of our world-systems shrinks. Then iii) the process may start again as people rebuild world-systems.

What is the long-term consequence of increasing world-system connectivity? Unfortunately, this means that there is a growing non-zero probability of a supersystem risk being realized over a shortening time horizon. Indeed, the impacts of anthropogenic climate change mean that in the coming decades, shocks will be continually experienced over a range of spatial and temporal scales. One only need look to the increased frequency of 100-year floods or record setting maximum daily temperatures from city to city around the world. Eventually, a supersystem risk will be realized of such magnitude that the connectivity of our world-systems may actually shrink. In some way, the 2020 pandemic is one such example: for instance, global air travel has greatly reduced, and it may never recover to what it was. This whole process — increasing connectivity to maximize economic efficiency, the realization of a supersystem risk and an eventual reduction in connectivity due to the impacts of the supersystem risk — resembles a process of Self-Organized Criticality35 (see Fig. 1 for a visual representation of this process). Self-Organized Criticality describes why some systems are attracted to catastrophe, and it has helped us understand the frequency and magnitude of forest fires, earthquakes and financial crashes to name a few examples. Perhaps a terrifying property of reflexive modernization is that the incentives we have created for ourselves will increasingly draw us towards creating supersystem risks17, until the point at which catastrophic collapse is inevitable. Perhaps the Anthropocene will be short-lived36.

Figure 1. Schematic showing the possible growth and eventual collapse of world-system connectivity. World-systems – social, transport, financial systems for example – are identified by the colored modules in the different networks. Initially these world-systems include connectivity within themselves, but, i) over time (i.e. decades, centuries) these world-systems have become highly connected, driven by incentives for economic efficiency. With greater connectivity comes supersystem risk and ii) at some point a super system risk is realized, and the connectivity of our world-systems shrinks. Then iii) the process may start again as people rebuild world-systems.

So What?

If the Anthropocene is on a collision course with catastrophe, driven by economic incentives for greater connectivity, what can be done? First, we must recognize that for a supersystem to be resilient, and indeed anti-fragile37, then it needs to learn from the shocks it experiences7. However, owing to our preoccupation with the future and with risk, we have actually become very good at preventing shocks. Even though we have and continue to experience shocks, it has been argued (albeit in a highly specific manner) that since the end of World War II, even though we have seen the growing threat of weapons of mass destruction38 and an increase in income inequality around the world39, we have actually experienced an unprecedented period of international peace and economic growth40. But, just like forests and rangelands that naturally experience regenerative wildfires, our world-systems must use shocks as opportunities to transform (see Fig. 2 for a depiction of the tradeoff in economic efficiency and resilience, and how anti-fragility manifests). Institutions that are not “fit” in the face of these socially created shocks must adapt and/or transform to be resilient to them in the future, and not persist through temporary fixes. For example, many of the financial institutions that created the 2008 crisis were bailed out (i.e. in the US, the Emergency Economic Stabilization Act of 2008). These short-term fixes promote maladaptation and ultimately lead to the recurrence of certain types of supersystem risks. Similarly, the 2020 response to the Covid-19 pandemic revealed differences in nations’ abilities to deal with the collective action challenge of limiting the spread of the virus through economic closures and social distancing; are we ready to learn from these experiences in the case of another pandemic in the (near) future? It is perhaps unwelcome to imagine shocks such as the 2008 financial crisis or the 2020 pandemic as being good for our global supersystem. But there can be many great lessons to be learned from them, that ultimately improve the functioning of our world-systems so that the impacts of future crises are less severe. Resilience in our world-systems is possible. The socio-economic systems we are embedded in are continually changing, and there is a wide spectrum of possible configurations that better incentivize local/regional socio-environmental resilience over global economic efficiency41,42 and the production of wealth. If we can recognize that the economic efficiency that currently provides us easy access to goods, services and jobs is also constructing supersystem risks, then we may be able to choose social, political and economic institutions that better serve the long-term prosperity of an Anthropocene that is socially and environmentally sustainable.

JRW would like to acknowledge funding from the DARPA grant HR00112020027. All authors would like to acknowledge conversations with Sam Bell, Robert Kennedy, David Wrathall and Demian Hommel as a major source of ideas for this article.

James R. Watson is an Assistant Professor in the College of Earth, Ocean and Atmospheric Sciences at Oregon State University. James is addicted to solving problems related to climate change risk management, complex systems science, sustainability, ecosystems and human behavior. Solutions he helps create take inspiration from the study of complex adaptive systems, leveraging mathematical theory, computational simulation and (big and small) data analytics. His interest extends to financial systems, the vertebrate immune system, housing markets, and sports data analytics. James received a B.Sc. in Biochemistry from Bristol University, a M.Sc. in Oceanography from the National Oceanography Centre, and a Ph.D. in Marine Science from the University of California Santa Barbara. James has also spent time researching in the Department of Ecology and Evolutionary Biology at Princeton University, and the Stockholm Resilience Centre.

Jamon Van Den Hoek is an Assistant Professor of Geography at Oregon State University where he leads the Conflict Ecology lab. His research seeks new insights on the agency, decision-making processes, and survival of refugees, internally displaced peoples, and others affected by violent conflict. He maps settlements, forests, and farms using satellite data to connect patterns of long-term landscape change to processes of conflict, displacement, resilience, and peace. Jamon was a NASA Postdoctoral Fellow at NASA Goddard Space Flight Center and completed his PhD in Geography at the University of Wisconsin-Madison where he was a National Science Foundation IGERT Fellow.

Laura E. R. Peters is a Postdoctoral Research Fellow at University College London cross appointed to the Institute for Risk and Disaster Reduction and the Institute for Global Health. Her research explores the complex interlinkages between natural hazard-related disasters and climate change, peace and conflict, and human health and wellbeing as long-term social and environmental processes. Working across dozens of international case studies, Laura investigates how heterogeneous – and at times deeply divided – societies build knowledge about, cope with, and act upon contemporary social and environmental changes and challenges, including those related to climate change and violent conflict. Laura completed her PhD in Geography at Oregon State University, her MA in International Peace and Conflict Resolution at American University, and her second MA in International Development and Cooperation at Korea University. Laura has seven years of applied career experience working for international non-governmental organizations and multilateral organizations based in Washington, D.C.