1. Introduction: from perfect storms to polycrises
As war, extreme weather, hunger, energy scarcity, inflation, pandemics, and myriad other calamities fill our daily news feeds, political leaders often declare that humanity is facing a ‘perfect storm’ of crises. This metaphor, however, is misleading (Homer-Dixon & Rockström, Reference Homer-Dixon and Rockström2022). It implies the current confluence of unfortunate events is merely a temporary coincidence – just plain bad luck.
But many of these leaders also recognize that today's crises are intertwined in vital ways (e.g. Georgieva, Reference Georgieva2022; Malpass, Reference Malpass2022): one crisis often seems to trigger or worsen another, which then triggers or worsens yet another; and interacting crises can produce impacts that are both different from and worse than the harms the crises would have produced separately. These leaders seem to intuit that the world's conjoined crises must be understood and addressed as a whole.
The term ‘polycrisis’ captures this intuition. It is being used by a growing number of commentators (Summers & Ahmed, Reference Summers and Ahmed2022; Wolf, Reference Wolf2022), international agencies (UNDP, 2022; UNICEF, 2023; WEF, 2023), policymakers (Juncker, Reference Juncker2018), and scholars (Davies & Hobson, Reference Davies and Hobson2022; Tooze, Reference Tooze2021). Yet the term remains underspecified – a buzzword with little substantive content. It is not yet associated with a rigorous field of inquiry that includes a framework of precisely defined core concepts and research heuristics that can sustain disciplined knowledge cumulation (Lakatos & Musgrave, Reference Lakatos and Musgrave1970). Without these elements, ‘polycrisis’ adds little to our understanding; but with these elements, the concept could help scholars generate actionable insights into the world's interwoven crises.
In this article, we provide the polycrisis concept with substantive content. We develop a research agenda for studying the causal mechanisms that entangle multiple global systems and that appear to be generating near-simultaneous global crises. We argue that a better understanding of humanity's predicament as a polycrisis can help the world address its interconnected challenges.
In Section 2, we define ‘polycrisis’ and highlight the concept's value in comparison with more familiar concepts. In Section 3, we argue that humanity is facing a global polycrisis; though not our first, it is unprecedented in crucial respects that we have yet to fully comprehend. Section 4 uses models from the complexity and sustainability literatures to identify several causal mechanisms likely operating among global crises today. It introduces two examples to illustrate these mechanisms: first, the cascading impacts of interactions between the Covid-19 pandemic, the Ukraine-Russia war, and climate change; and, second, the potentially reinforcing feedbacks between economic turmoil, nationalist authoritarianism, and declining international cooperation that could tip the world into mass violence. In the concluding section, we summarize some of this nascent field's key insights and policy implications, while identifying future research directions.
2. What is a (global) polycrisis?
Complexity theorists Edgar Morin and Anne Brigitte Kern coined the term ‘polycrisis’ over two decades ago. They argued that the most ‘vital’ problem of the day was not any single threat but the ‘complex intersolidarity of problems, antagonisms, crises, uncontrollable processes, and the general crisis of the planet’ (Morin & Kern, Reference Morin and Kern1999, p. 74). More recently, sustainability scholar Mark Swilling (Reference Swilling2013, Reference Swilling2020) used ‘polycrisis’ to capture the complex interactions between crises in the global political economy that multiply those crises' overall impact. In the 2010s, European scholars and leaders (most notably then-President of the European Commission Jean-Claude Juncker) adopted the term to label the simultaneous migration, financial, and Brexit crises afflicting Europe (Juncker, Reference Juncker2018; Zeitlin et al., Reference Zeitlin, Nicoli and Laffan2019). And in the months following Russia's invasion of Ukraine in February 2022, Columbia University's Adam Tooze and researchers at the Cascade Institute used ‘polycrisis’ to characterize the complex interactions between the effects of the war, climate change, and the pandemic (Lawrence et al., Reference Lawrence, Janzwood and Homer-Dixon2022; Tooze, Reference Tooze2022).
But it was only at the World Economic Forum's annual meeting in Davos in January 2023 that the polycrisis idea gained wide currency among commentators, policymakers, and business elites (Serhan, Reference Serhan2023). This surge in use engendered broad criticism of the concept and, unfortunately, more confusion than clarity (Homer-Dixon et al., Reference Homer-Dixon, Lawrence and Janzwood2023).
Some critics argue that the polycrisis idea obscures the operation of capitalist interests that are at the root of the world's woes (Sial, Reference Sial2023); they associate the term with ‘Davos elites’ and their supposed faults. Others argue that our present predicament is not truly novel; the world has seen intersecting crises before, so we do not need a new concept to describe our situation today (Kluth, Reference Kluth2023). And at least one International Relations scholar has muddied the waters by misrepresenting polycrisis arguments as ‘neo-Malthusian’ – that is, explanations of complex social phenomena that overemphasize the causal role of population growth and resource depletion (Drezner, Reference Drezner2023).
Neologisms always provoke contention. But the disputes in this case risk distracting us from a core (and presumably shared) goal: to better understand and address our world's very real crises. We believe the polycrisis concept – if defined clearly and translated into a productive program of research and action – can help us pursue this goal.
In this spirit, we define a ‘global polycrisis’ as the causal entanglement of crises in multiple global systems in ways that significantly degrade humanity's prospects (Lawrence et al., Reference Lawrence, Janzwood and Homer-Dixon2022). We unpack this definition by first defining ‘crisis’, then by identifying interactions among crises that constitute a ‘global polycrisis’, and finally by distinguishing this latter term from related concepts of ‘systemic risk’, ‘catastrophic risk’, and ‘existential risk’.
We define a crisis as a sudden (non-linear) event or series of events that significantly harms, in a relatively short period of time, the wellbeing of a large number of people (Homer-Dixon et al., Reference Homer-Dixon, Walker, Biggs, Crépin, Folke, Lambin, Peterson, Rockström, Scheffer, Steffen and Troell2015).Footnote i
More colloquially, it is an extremely harmful emergency that requires urgent response lest even greater harm ensue. This definition diverges slightly from early and modern understandings of the term: for ancient Greeks, a crisis was the decisive moment at which an illness veers toward death or recovery; in modern politics, it is an alarming situation that could steer the course of history and therefore demands rapid resolution (Koselleck, Reference Koselleck2006). Both early and modern usages reference a rupture of normalcy that has fateful consequences and thus requires decisive action. Modern usage also highlights epochal change over time (in ways that resonate with our discussion of system stresses in Section 4 below). In contrast, our definition of crisis emphasizes immediate harms.
Our definition of ‘crisis’ is precise enough to support development of objective criteria of crisis occurrence and severity. Such criteria could make it harder to use the term selectively and inconsistently to emphasize some problems and solutions over others and thereby to serve particular interests. Declaration that a crisis is occurring is often a key step in the securitization of an issue: a problem like cross-border migration or climate change becomes a crisis, and thus a matter of national security not because of its inherent features, but because certain actors convince relevant audiences (generally policymakers) that the issue constitutes an existential threat to the nation and therefore requires responses outside the realm of normal politics (Buzan et al., Reference Buzan, Wæver and de Wilde1998, pp. 21–47). Any definition of crisis will have political implications, but objective criteria (to the extent they can be developed) help limit politicized manipulation of the term. By referring to facts and evidence about actual – rather than counterfactual – occurrences, our definition helps to narrow the scope of what can be credibly and consistently labeled a crisis.
If a crisis is an extremely harmful emergency, then the poly- in polycrisis denotes multiple such events. But this prefix is of little use if it denotes any coincidence of crises or simply refers to all the world's ills.Footnote ii On this point, the concept's critics are correct. We therefore emphasize crises that are causally inter-related with one another, and we draw upon the systemic risk literature and systems thinking more broadly to discern the types of crisis connections that constitute a polycrisis.
Conventional risk assessment focuses on the likelihood and potential harm of particular events such as a car accident, fire, or bankruptcy. In contrast, systemic risk assessment focuses on ‘the risk or probability of breakdowns in an entire system, as opposed to breakdowns in individual parts or components, [as] evidenced by co-movements (correlations) among most or all parts’ (Kaufman & Scott, Reference Kaufman and Scott2003, p. 371). Our elaboration of the polycrisis concept here adopts two core implications of this systemic risk idea (Renn, Reference Renn2016; Renn et al., Reference Renn, Lucas, Haas and Jaeger2019; Schweizer, Reference Schweizer2021):
(1) Intra-systemic impact: A disruption that affects one part or area of a single system quickly spreads to disturb the entire system (via multiple, ramifying chains of cause and effect, or some form of contagion, through the system's causal network).
(2) Inter-systemic impact: The disruption of the initial system may spill outside that system's boundaries to disrupt other systems.
The concept of systemic risk ‘assumes a systems perspective’ (Schweizer, Reference Schweizer2021, p. 79). It presupposes that ‘connections between elements of the system’ are sufficiently dense that a single disruption can sometimes generate ramifying impacts throughout the system. It also implies that discernable boundaries separate one system from another (Figure 1), although discrete systems may influence each other by exchanging energy, matter, information, and biota (Box 1).
At a rudimentary level, four vectors can carry a crisis within and across systems and from one part of the world to another, thereby inflicting significant harms:
• Energy, such as the kinetic energy generated by earthquakes and hurricanes.
• Matter, such as the toxins and pollutants that harm organisms and ecosystems.
• Information, consisting of instructions and symbolic representations – including genetic and digital codes, news feeds, ideologies, money, policies, and laws – that can be communicated between agents.
• Biota, such as viruses, bacteria, and other organisms that can disrupt the biological and physiological functions of other organisms. (This category may be considered a special combination of energy, matter, and information that involves lifeforms.)
Any given crisis event will likely feature some combination of these vectors, whether simultaneously or sequentially. A hurricane, for example, disperses kinetic energy through wind and rain, which can cause matter in the form of floodwater to inundate populated areas and create conditions for the spread of pathogenic biota, while information about the disaster may provoke panicked, inappropriate responses. Crises may stem either from vectors that carry harms or from sudden disruptions of vectors that carry necessities, as when energy outages leave households vulnerable to harsh winters. Social power can be understood as an actor's ability to manipulate these vectors to get another actor to do what they otherwise would not do (Dahl, Reference Dahl1957), in ways that can create a crisis by intention, negligence, or accident.
Today's planet-spanning webs of connections – including those arising from Earth's biophysical features and others produced by humanity's globalized economic activity – provide the conduits through which these vectors travel around the globe. This web of connections includes our societies' telecommunication networks; pipeline networks and electrical grids; roads, canals, and air and shipping routes; supply chains and trade, finance, and monetary systems; and links among elements of Earth's climate and ecological systems.
Our polycrisis concept similarly assumes that initially limited disruptions can affect an entire system and then spread to other systems. But it differs from the systemic risk concept in three important ways. First, whereas the ultimate referent of the systemic risk concept (and of the risk concept more generally) is the potential harm that might arise, the referent of the polycrisis concept is the realization (or activation) of chains of cause and effect that cause harms. Second, a systemic risk is generally assumed to arise from just one or two systems, but a polycrisis (by definition) arises from interactions among multiple systems.Footnote iii And finally, whereas the systemic risk literature highlights the complexity of risks themselves, our approach to polycrisis instead emphasizes the complexity of the systems in which the risks develop. This complexity creates the possibility for systemic failure and inter-systemic effects; that is to say, systemic complexity creates systemic risks (Goldin & Mariathasan, Reference Goldin and Mariathasan2016; Nyström et al., Reference Nyström, Jouffray, Norström, Crona, Søgaard Jørgensen, Carpenter, Bodin, Galez and Folke2019). Box 2 presents the key features of global systems that enable polycrises to develop and grow.
Operating together, the vectors and conduits described in Box 1 create highly complex global systems. These systems exhibit five key properties that help generate polycrises while hampering crisis mitigation.
• Multiple causes: The operation of many causes simultaneously makes cause and effect relationships difficult to trace and presents decisionmakers with acute policy trade-offs. Causes may also interact synergistically so that their combined effects are qualitatively different than the sum of effects they would have separately (Jervis, Reference Jervis1997).
• Non-linearity: Complex global systems exhibit nonlinear behavior – that is, perturbations of such systems can produce disproportionately large (or small) changes in the system's behavior. An important source of nonlinearity is the existence of multiple stable states or equilibria that are separated by thresholds. A system can flip from one equilibrium to another (a critical transition or tipping event) when feedbacks in key processes that sustain the system's equilibrium shift from negative to positive – i.e., from self-dampening to self-reinforcing causal loops (Scheffer, Reference Scheffer2009). Tipping events can also result from interactions between adjacent systems (Rocha et al., Reference Rocha, Peterson, Bodin and Levin2018).
• Hysteresis: System flips are generally not reversible; a return to the previous system equilibrium is often impossible.
• Boundary permeability: Casual processes operate on multiple time scales within and among natural, social, and technological systems; they cross boundaries of administrative and political units and social sectors, while requiring integrated knowledge from diverse scientific disciplines.
• ‘Black swan’ outcomes: The probability density functions describing the distribution of events generated by complex global systems are rarely normal (i.e. Gaussian); they often have long tails, indicating a non-negligible risk of extreme outcomes. Leaders, in contrast, face institutional pressures to concentrate on immediate and probable risks.
These five properties create deep uncertainty that profoundly hinders effective management of outcomes. Multiple causes and nonlinearity weaken decisionmakers' ability to predict which policy changes will matter when. Tipping events and hysteresis undermine trial-and-error learning; a maladaptive behavior can generate benefits until a threshold is crossed, at which point costs are unavoidable, damage irreversible, and any learning too late. Ineffective learning then lowers the public's willingness to accept costs to lessen risk (Barrett & Dannenberg, Reference Barrett and Dannenberg2014).
Because risks arising within complex global systems tend to transcend administrative, social, and scientific boundaries, they often exceed managers' professional expertise and are consequently downplayed or even ignored. And when crises affect multiple administrative and political domains, actors may choose to free ride on others' investments in solutions. Finally, deep uncertainty fosters competing policy prescriptions, aggravating a pernicious loss of trust in governments' problem-solving capacity. In some cases, uncertainty can be reduced; in others, it is either practically or intrinsically irreducible (Janzwood, Reference Janzwood2022; Walker et al., Reference Walker, Harremoës, Rotmans, van der Sluijs, van Asselt, Janssen and von Kraus2003).
By focusing on crises within and across systems, our approach highlights a crucial feature of polycrises: that the conjoined harms of multiple crises are different from, and generally worse than, the harms each crisis would produce in isolation, were their host systems not so deeply interconnected (Lawrence et al., Reference Lawrence, Janzwood and Homer-Dixon2022, p. 2). What may appear to be separate crises in different systems in fact exacerbate and reshape one another to form a conjoined polycrisis that must be understood and addressed as a whole. In the language of complexity scientists, a polycrisis is an emergent phenomenon.
Like systemic risk (ISC et al., 2022), a polycrisis can occur at different scales – local, national, regional, or global – indeed at any scale that hosts interacting systems. Here, however, we are particularly concerned with crises interacting at the global scale, with a spatial extent that affects the whole planet and/or all of humanity.Footnote iv Global polycrises (and global systemic risks) arise from the organization of human activities into complex global systems (as defined in Figure 1) structured in ways that enable disruptions to spread quickly around the world (as outlined in Box 2).
In the interest of establishing a research agenda, we have adopted a harm threshold that remains somewhat ambiguous and hence leaves room for future refinements. In the extreme, a polycrisis could reach the severity of a ‘catastrophic risk’, an event that kills 10–25% of humanity (Cotton-Barratt et al., Reference Cotton-Barratt, Farquhar, Halstead, Schubert and Snyder-Beattie2016; Kemp et al., Reference Kemp, Xu, Depledge, Ebi, Gibbins, Kohler, Rockström, Scheffer, Schellnhuber, Steffen and Lenton2022) or brings about the collapse of human civilization (GCRI, 2023). It could even become an ‘existential risk’ that extinguishes humanity entirely. But a polycrisis, by our definition, does not need to reach these levels of harm; and, in contrast to accounts of individual existential and catastrophic threats (arising from, for instance, an asteroid hitting Earth), a polycrisis necessarily involves multiple crisis events. It could involve massive immediate casualties, but also a widespread and sustained decline in the quality of life into the future.
Based on these considerations, we define a global polycrisis as the causal entanglement of crises in multiple global systems in ways that significantly degrade humanity's prospects. The causal interactions between constituent crises are significant enough to produce emergent harms that are different from, and usually greater than, the sum of the harms they would produce separately. Consequently, these crises must be addressed as a whole; they cannot be resolved individually. While our approach to polycrisis incorporates key aspects of other definitions, it is specifically intended to aid scientific research into the nature of polycrisis by emphasizing the causal interactions that connect global systems and spread crises among them. Our definition relates to other important concepts (such as systemic risk) but adds essential novelty by highlighting the causal entanglement of multiple crises – interconnections that abound but remain sparsely understood, as explained in the sections below.
3. Are we in a global polycrisis?
We argue here that the world is currently experiencing a global polycrisis and that this situation is worsening. Constituent crises include the lingering health, social, and economic effects of the Covid-19 pandemic; stagflation (a persistent combination of inflation and low growth); volatility in global food and energy markets; geopolitical conflict, especially between assertive authoritarian regimes (including China and Russia) and the democratic West, which is leading to a partial decoupling of American and Chinese economies; political instability and civil unrest in countries both rich and poor arising from economic insecurity, ideological extremism, political polarization, and declining institutional legitimacy; and increasingly frequent and devastating weather events generated by climate heating. These crises are destroying livelihoods and lives around the globe and are undoubtedly diminishing humanity's prospects. Moreover, they are certainly interconnected, although exactly how remains unclear.
This is not humanity's first polycrisis. We experienced at least two additional instances in the last half century, though some may argue they were not truly global. The oil shocks of the 1970s arose from conflicts in the Middle East and generated severe international energy shortages that contributed to, and interacted with, stagflation in the world economy (Progressive International, 2023). The 2008–09 global financial crisis intersected with oil supply constraints and long-term stresses in food production to produce cascading bankruptcies, food price hikes, and political unrest worldwide (Biggs et al., Reference Biggs, Biggs, Dakos, Scholes and Schoon2011; Homer-Dixon et al., Reference Homer-Dixon, Walker, Biggs, Crépin, Folke, Lambin, Peterson, Rockström, Scheffer, Steffen and Troell2015).Footnote v
While the present polycrisis features some of the same constituent crises – including energy and food shocks, stagflation, and financial instability – it is unprecedented in crucial ways (Homer-Dixon, Reference Homer-Dixon2023; Lähde, Reference Lähde2023). First, the world is far more interconnected now than it was during the OPEC oil shocks. Between 1980 and 2020, air freight increased sixfold to 180 billion-ton-kilometers per year, the number of air passengers nearly tripled to 1.8 billion annually, and internet usage increased from virtually 0 to 60% of the world's population. Meanwhile, the total value of world merchandise trade increased twelve-fold between 1980 and 2022 to nearly 25 trillion US dollars annually (at current prices), and container port traffic has more than tripled since 2000 to almost 800 million 20-foot-equivalent-units in 2020.Footnote vi
The ‘conduits’ of this extreme connectivity – aircraft, container carriers, fiber-optic cables, and the like – now carry immense circum-planetary flows of the ‘vectors’ of matter, energy, biota, and information (Box 1). The conduits also create and sustain multi-continental markets and globalized corporations that in turn encourage increasing standardization and homogenization among system elements, from financial instruments to germ plasm for agricultural goods to computer operating systems and social media platforms. This homogenization then enables even denser interconnection, in a powerful positive feedback.
Unfortunately, complex systems that feature both high connectivity and high homogeneity among system elements can be especially prone to rapid, discontinuous change (Scheffer et al., Reference Scheffer, Carpenter, Lenton, Bascompte, Brock, Dakos, van de Koppel, van de Leemput, Levin, van Nes, Pascual and Vandermeer2012), much as closely planted agricultural monocrops are susceptible to devastation by pathogens. By striving to maximize efficiency and open access to markets while stripping away social and environmental safeguards, neoliberal arrangements have exacerbated both homogenization and hyper-connectivity in the global economy, generating recurrent crises and worsening stresses both in the economy (for instance, by increasing inequality) and in other systems (for instance, by damaging the ecosphere).
Even in the absence of high homogenization, gradual shifts in exogenous conditions can erode a highly connected system's resilience until its stabilizing feedbacks are overwhelmed, and it flips to a different equilibrium (Scheffer, Reference Scheffer2009). And systems that may be resilient on their own can become more vulnerable to such flips when they become tightly connected to other systems (Buldyrev et al., Reference Buldyrev, Parshani, Paul, Stanley and Havlin2010; Gao et al., Reference Gao, Liu, Li and Havlin2015); unexpected vulnerabilities can arise when system elements not designed to work together are inadvertently connected (Perrow, Reference Perrow1999).
In sum, the interlinked architecture of our global systems is at the heart of the current polycrisis, because it worsens risks as diverse as financial turmoil, pandemics, economic inequality, and ideological extremism (Centeno et al., Reference Centeno, Nag, Patterson, Shaver and Windawi2015; Helbing, Reference Helbing2013; Rodrik, Reference Rodrik2011). These systemic risks are ‘endemic to globalization’; they can be managed (by reforming the neoliberal economic order, for instance) but not eliminated (Goldin & Mariathasan, Reference Goldin and Mariathasan2016, p. xiii).
The present polycrisis is also unprecedented in a second respect. Human resource consumption and pollution output are pushing Earth's physical and ecological systems far from their previous equilibria, imperiling the stability of many other global systems critical to human wellbeing, from food production to international security. For instance, our emissions of greenhouse gases have created an energy imbalance at the planet's surface (more heat coming in from space than going out) of about 1.36 Watts per square meter (Hansen et al., Reference Hansen, Sato, Simons, Nazarenko, Sangha, Kharecha, Zachos, von Schuckmann, Loeb, Osman, Jin, Tselioudis, Jeong, Lacis, Ruedy, Russell, Cao and Li2023). This extra energy – now equivalent to nearly one million Hiroshima-sized atomic bombs exploded in the atmosphere every day – is producing increasingly extreme storms, floods, heat waves, and droughts, affecting billions of people and worsening population displacement, social instability, and conflict (Adelphi & PIK, 2020; Ide et al., Reference Ide, Brzoska, Donges and Schleussner2020; Schleussner et al., Reference Schleussner, Donges, Donner and Schellnhuber2016).
Together, hyper-connectivity and the destabilization of ecospheric systems are amplifying and accelerating crisis events worldwide (Figure 2). For example, since HIV first appeared four decades ago, outbreaks of zoonotic viral disease have become increasingly severe and frequent, from the SARS outbreak of 2002 to H1N1 in 2009, MERS in 2012, Ebola in 2014, Zika in 2015, Ebola again in 2018, Covid-19 in 2019, and most recently mpox and Marburg (Araf et al., Reference Araf, Maliha, Zhai and Zheng2023; CFR, 2023; Smith et al., Reference Smith, Goldberg, Rosenthal, Carlson, Chen, Chen and Ramachandran2014). Meanwhile, climate heating is also accelerating: between 1970 and 2010, Earth's tropospheric temperature increased about 0.18 °C per decade; between 2010 and 2040, warming is predicted to increase to 0.27 °C per decade, a rise in rate of 50% (Hansen et al., Reference Hansen, Sato, Simons, Nazarenko, Sangha, Kharecha, Zachos, von Schuckmann, Loeb, Osman, Jin, Tselioudis, Jeong, Lacis, Ruedy, Russell, Cao and Li2023, p. 21). And because this warming makes zoonotic disease outbreaks more likely, two seemingly discrete crises – pandemics and calamitous weather – are becoming increasingly entwined (Carlson et al., Reference Carlson, Albery, Merow, Trisos, Zipfel, Eskew and Bansal2022).
But global crises are not just amplifying and accelerating, they also appear to be synchronizing. ‘We're seeing what occurs when everything happens everywhere all at once’, says International Relations theorist Stephen Walt (Reference Walt2022). Complex and largely unrecognized causal links among the world's economic, social, and ecological systems seem to be causing many risks to go critical at the same time or in quick succession (Figure 3). Indeed, ‘the failure to take into account feedbacks across systems’ is a crucial emerging risk itself (Future Earth, 2020, p. 6).
While scientific knowledge of individual systemic risks like climate change and zoonotic viral disease is often deep, our grasp of causal mechanisms linking these risks and the crises they generate remains shallow (ISC et al., 2022, p. 8). For instance, the World Economic Forum's annual Global Risk Report identifies apparent links among risks but does not examine amplifying feedbacks in detail. Below, therefore, we offer an analytical framework to help advance our understanding of the causal mechanisms driving the present polycrisis.
4. The causal mechanisms of crisis entanglement
‘Synchronization’ can mean several things. In physics, synchronization occurs when interactions between oscillating objects cause them to align their rhythms so that events happen at the same time or with the same periodicity (Pikovsky et al., Reference Pikovsky, Rosenblum and Kurths2007). Synchronization often homogenizes behavior by causing system elements to act in the same way, as when glow bugs flash in unison or investors all try to sell off bad stocks at the same time (Strogatz, Reference Strogatz2003). And the term synchronization may be used more loosely to refer to events that occur in quick succession.
The apparent synchronization of global crises (in any of the above senses) raises a crucial question: what sorts of interactions and feedbacks are aligning crises in multiple global systems? These relationships remain opaque and underexplored. We therefore propose an analytical framework to guide investigation of the causal mechanisms connecting global crises.
4.1 The basic model: crisis in a single system
Scholars and policymakers tend to silo their analyses of, and responses to, crises; that is, they tend to see the causes and effects of a given crisis through the lens of a single system. Such parsimony can be a useful analytical starting point. Beginning, therefore, with a single system, our basic model (Figure 4) proposes that a crisis occurs when one or more slow-moving stresses interact with one or more fast-moving trigger events to push the system out of its established equilibrium and into a state of disequilibrium or instability.Footnote vii In line with our earlier definition of crisis (Section 2), this disequilibrium manifests itself as a sudden (non-linear) event or series of events that significantly harms a large number of people.
A complex system is not static. Constantly operating internal processes (such as negative feedbacks) keep the system's state (represented in Figure 4b as a ball) within a certain range of values (depicted as a ‘basin of attraction’ in a ‘stability landscape’). Stresses are slow-moving processes – pressures, emerging contradictions, and deepening vulnerabilities – that accumulate in the system over time and weaken its stabilizing feedbacks, reducing their ability to hold the system's state within its established range.Footnote viii Metaphorically, the basin in which the system resides becomes shallower.
Stresses often operate at the global scale, and because they are slow-moving, their change over time is usually somewhat predictable. In global systems, stresses currently include growing socio-economic inequalities, increasing resource scarcities, economic over-leveraging, climate heating, and ecological degradation, among many others. By reshaping the stability landscape, these stresses shift the probabilities of future global developments and create systemic risks – that is, potential pathways across that landscape to crisis.
A trigger event is a fast-moving process that interacts with stresses to push a system state out of equilibrium. If stresses have made the system's basin of attraction shallower, a trigger event of a given magnitude will more easily cause such disequilibrium. Trigger events are usually stochastic, unpredictable, and local or regional in scale, but they have global-systemic consequences. They include phenomena like political uprisings, price spikes in critical goods and services, major corporate bankruptcies, and the loss of keystone species in specific ecosystems.
A system enters crisis when it leaves its established basin of attraction. A crisis thus has three defining properties: the system state is unstable (i.e. out of equilibrium), the change in system state occurs relatively suddenly, and the resulting instability causes significant human harm. Pushed from equilibrium, the system is in a turbulent state that disrupts stabilizing mechanisms and generates harmful outcomes, such as loss of income or deaths and injuries from violent conflict, malnutrition, starvation, or disease.
A crisis ends when the system returns to equilibrium – by either re-entering its original basin of attraction or moving to a new one. If the system state returns to its original basin and that basin remains shallow, a crisis will likely erupt again. If the system state settles into a new basin of attraction, it has completed a critical transition; it has flipped from one set of system behaviors to another with its own stabilizing internal processes.
Ideally, a crisis ends with the system entering a basin that reinforces normatively beneficial system behaviors and which is sufficiently deep (i.e. stable) to prevent another crisis. But the system could also enter a harmful and undesirable – but still highly stable – basin, perhaps one with widespread economic deprivation and political repression. In these circumstances, it is the system's newfound stability in a pernicious state, rather than its crisis instability, that creates significant harm.Footnote ix For example, systems such as slavery and imperialism caused immense suffering over long historical periods – not as crises, but due to their lamentable stability and resilience.
The global financial crisis of 2008–9 illustrates our basic model. It arose from the conjunction of several slow-process stresses, including growing worldwide trade in opaque financial instruments securitized by overvalued housing markets, and tightening balance-sheet interdependencies among major financial institutions stemming from cross-ownership of these instruments. The collapse of Lehman Brothers was the trigger event that started a cascade of defaults. The crisis ended when central banks rescued major commercial banks from default, slashed interest rates, and injected unprecedented amounts of liquidity into national economies. The global economic system settled into a new disinflationary equilibrium of weak demand, low growth, and exceptionally low interest rates that lasted until the Covid-19 pandemic.
Through this entire period and up to the present, the global economy has continued to experience additional powerful stresses – including rising economic inequality within most nations and worsening global heating – that have progressively weakened its social and ecological foundations and contributed to a long-term fall in the secular rate of global economic growth (Homer-Dixon, Reference Homer-Dixon2020, p. 204). These (and other – see e.g. Roubini, Reference Roubini2022) changes amount to a steady shallowing of global capitalism's basin of attraction that is boosting the risk of future systemic crises.
No conceptual schema can fully capture the intricate causal, spatial, and temporal features of specific global crises. But our basic single-system model should help researchers distinguish between the three core elements of stress, trigger, and crisis and then map interactions among these elements. Figure 5 shows possible types of within-system interaction.
4.2 Crisis interaction between multiple systems
A global polycrisis, however, is characterized by relationships between systems. In Figure 6, we show how the elements of our basic model (stresses, triggers, and crises) can interact among multiple systems.
The possible inter-systemic interactions shown in Figure 6 draw upon – and echo – advances in ecological research. Just as other scholars and policy makers tend to address crises in single systems, ecologists have largely studied critical transitions in isolated ecosystems. But recent, leading-edge work in ecology identifies causal relationships between such transitions in multiple ecosystems (Keys et al., Reference Keys, Galaz, Dyer, Matthews, Folke, Nyström and Cornell2019; Klose et al., Reference Klose, Wunderling, Winkelmann and Donges2021; Rocha et al., Reference Rocha, Peterson, Bodin and Levin2018).
Rocha et al. (Reference Rocha, Peterson, Bodin and Levin2018) compare the thirty ecosystem critical transitions mapped in the Regime Shifts DatabaseFootnote x and identify three broad types of causal relationships between them:Footnote xi
• Common stresses: A common stress may weaken the resilience of multiple systems, or the stresses affecting one system may interact with stresses in another, as depicted in Figure 6a.
• Domino effects: A crisis in one system may affect the stresses in another system, cause a triggering event that pushes another system into crisis, or reshape a crisis in another system, as depicted in Figures 6e and 6f. Domino effects operate in temporal sequence.
• Inter-systemic feedbacks: Stresses, trigger events, and other events generated by a crisis can form feedback loops that either dampen or, more commonly, escalate crises in two or more systems. Feedback effects can be depicted by combinations of the processes shown in Figure 6, as illustrated in Figure 7.
We propose additional possibilities: ‘common triggers’ by which the same event can activate crises in multiple systems (Figure 6b), as well as possible causal interactions between stresses in different system (Figure 6c) and stresses in one system that generate trigger events in another system (Figure 6d). All six forms of interaction depicted in Figure 6 can be thought of as ideal types; together they provide a ‘grammar’ of causal interactions between systems that can be used to develop hypotheses in polycrisis research. The remainder of this section provides further applications and examples.
4.3 Common stresses and systemic synchronization
Many global systems are currently undergoing radical change; this simultaneity of change is probably not coincidental. It suggests common stresses are causing the synchronization of underlying system behavior (Figure 6a), which may account (at least in part) for the acceleration, amplification, and apparent synchronization of today's global crises.
• The Earth environmental system is leaving its Holocene equilibrium and entering a period of instability due to anthropogenic perturbation of the climate and other physical and ecological systems (Armstrong McKay et al., Reference Armstrong McKay, Staal, Abrams, Winkelmann, Sakschewski, Loriani, Fetzer, Cornell, Rockström and Lenton2022; Barnosky et al., Reference Barnosky, Hadly, Bascompte, Berlow, Brown, Fortelius, Getz, Harte, Hastings, Marquet, Martinez, Mooers, Roopnarine, Vermeij, Williams, Gillespie, Kitzes, Marshall, Matzke and Smith2012; Rockström et al., Reference Rockström, Gupta, Lenton, Qin, Lade, Abrams, Jacobson, Rocha, Zimm, Bai, Bala, Bringezu, Broadgate, Bunn, DeClerck, Ebi, Gong, Gordon, Kanie and Winkelmann2021; Steffen et al., Reference Steffen, Rockström, Richardson, Lenton, Folke, Liverman, Summerhayes, Barnosky, Cornell, Crucifix, Donges, Fetzer, Lade, Scheffer, Winkelmann and Schellnhuber2018). This instability is already causing enormous human harm, and its effects could become catastrophic in the near future (Kemp et al., Reference Kemp, Xu, Depledge, Ebi, Gibbins, Kohler, Rockström, Scheffer, Schellnhuber, Steffen and Lenton2022; Xu et al., Reference Xu, Kohler, Lenton, Svenning and Scheffer2020).
• The global human energy system has begun to shift away from its dependence on fossil fuels. Whether this shift will culminate in a new zero-carbon energy equilibrium is uncertain: technological bottlenecks and incumbent opposition may block its progress. The shift's economic benefits are also uncertain: it could ultimately force humanity to decrease its energy consumption per capita (Hall, Reference Hall2018; Smil, Reference Smil2022).
• The international security system is changing from a world order based on American leadership (a ‘pax Americana’) toward an uncertain and likely less-stable multipolar order defined by the rise of China and the diffusion of power to a much wider range of actors (Gilpin, Reference Gilpin2002; Ikenberry, Reference Ikenberry2014; Nye, Reference Nye2011). Historically, such transitions have been accompanied more often than not by major war (Allison, Reference Allison2017; Gilpin, Reference Gilpin1988).
• The global economic system is shifting from a neoliberal economic regime – one undermining itself through worsening instability, inequality, and ecospheric externalities – to a yet indeterminate regime, but one likely involving increased dirigisme and economic integration within ideological blocs (Birdsall & Fukuyama, Reference Birdsall and Fukuyama2011; Monbiot, Reference Monbiot2016; Rodrik, Reference Rodrik2011; Rodrik, Reference Rodrik2019).
• The information system is being revolutionized by artificial intelligence, with unclear but likely unprecedented implications for employment, decision making, and personal, national, and global security.
The simultaneity of radical change across these systems likely arises, in significant part, from their interdependence, as we argued in section 3 above. Stresses affecting one system can create (or constitute) stresses in others (Figures 6a and 6c). Stresses in the global energy system, for example, include the declining thermodynamic quality of remaining fossil fuel deposits, a trend that increases the energy cost (and therefore carbon emissions) of extraction. Fossil fuel emissions then create stresses in the Earth system, such as climate heating and ecosystem disruption. But possibilities for substituting other, zero-carbon energy sources remain limited (Hall, Reference Hall2018). Most alternatives, for instance, have relatively low power density, which makes them ill-suited as primary energy sources for today's high power-density-of-consumption urban regions and manufacturing facilities (Smil, Reference Smil2016). Fossil fuels also still provide energy for nearly all long-distance transportation and remain essential to steel, cement, plastic, and fertilizer production (Smil, Reference Smil2022, pp. 76–102). Stresses in the global energy system thus create stresses in global food, transportation, and economic systems.
Additionally, stresses in one global system can stimulate or constrain reorganization in others. For example, the Earth system's post-Holocene transformation is influencing change in the global energy system and thereby the global economic system. Hegemonic competition in the international security system could reduce governmental collaboration to reorganize the global energy system so as to reduce, in turn, that system's impacts on the Earth system.
A framework called adaptive cycle theory suggests that a number of global systems may be on the cusp of catastrophic reorganization. Global energy, food, and financial systems have become increasingly complex, and their sub-components increasingly specialized and connected, as firms have competed to maximize productivity and efficiency. These changes have made these systems more rigid and less resilient in some respects. Systems exhibiting such characteristics, adaptive cycle theory argues, are susceptible to breakdown and reorganization (Gunderson & Holling, Reference Gunderson and Holling2002; Holling, Reference Holling2001). When multiple systems align at this phase of the cycle – as several global systems appear to be doing now – breakdown in one may trigger breakdowns in others.
4.4 Domino effects between global systems
Such a cascade of breakdowns across systems would be an example of domino effects. The domino metaphor implies a linear chain of cause and effect, in which one crisis causes another, and so on. The interactions between global crises are, of course, not so simple. Stresses and triggers can interact across systems (Figures 6c and 6d); a crisis in one system may affect the stresses and/or the trigger events that push another system into crisis (Figure 6e); and the events generated by one crisis may influence the behavior of another system in crisis (Figure 6f). These types of interactions combine across multiple systems to form multicausal networks, in contrast to simple causal chains.
Figure 8 illustrates domino effects by mapping a causal network of stresses, triggers, and crises among several global systems – specifically, the health, environmental, economic, transportation, international security, and social order and governance systems – from the past through the present to possible (and somewhat speculative) outcomes in the future. The left-to-right temporal logic of such maps helpfully traces the course of events, but it cannot capture the recursive feedback loops that powerfully drive synchronization. Those feedbacks are illustrated instead by the causal loop diagrams in Figure 9.
4.5 Inter-systemic feedback loops
Domino effects are one-way causal relationships. But system behaviors can sometimes influence their own causes, creating feedback loops. Negative (i.e. dampening) feedbacks tend to stabilize systems by counteracting change, such as when markets correct for overvalued assets. Positive (i.e. self-amplifying) feedbacks involve two or more variables that intensify one another in spirals of run-away growth or decay, such as arms races or stock market crashes.
We argue that feedbacks arise from combinations of the interactions depicted in Figure 6 and produce the crisis synchronization manifested in a polycrisis. Although one crisis may on occasion dampen another – as when, for example, a stock market crash produces a communication system outage that slows herd behavior – the real danger arises when interactions among two crises' causes and effects create a positive feedback in which each crisis keeps worsening the other. Positive feedbacks can quickly overwhelm institutional safeguards and controls. And they can create an acute policymaking dilemma in which one crisis cannot be resolved without remediating a second one – but the second cannot be resolved without remediating the first.
Figure 9 illustrates several harmful positive feedbacks that appear to be forming today within and between the global systems identified in Figure 1. Compared to Figure 8, which shows how stresses, triggers, and crises can cascade unidirectionally over time, Figure 9 illustrates the back-and-forth (or cyclical) interactions between crises, triggers, and stresses.
In Figure 9a, economic turmoil arising, for instance, from inflation, financial crisis, and debt – or perhaps due to scarcities of key resources such as energy, food, water, and raw materials – creates mass grievances and institutional opportunities for populist leaders to capture political power and weaken the rule of law. These leaders' actions to establish authoritarian regimes simultaneously draw on and amplify nationalist, chauvinistic, and anti-globalization ideologies, often by scapegoating foreigners, cosmopolitan elites, and internal minorities. Although their efforts to decouple the national economy from the world economy generally worsen internal economic turmoil, this turmoil, paradoxically, often exacerbates the grievances and opportunities the leaders can exploit to consolidate their power (by blaming ‘foreign elements’ or ‘internal enemies’ for the economic crisis). In the last decade, this feedback has operated in such diverse countries as Venezuela, Nicaragua, Russia, Turkey, Zimbabwe, Myanmar, and Sri Lanka.
In Figure 9b, we show that populist authoritarian regimes espousing nationalist and anti-globalization ideologies generally decrease their participation in international institutions, reduce their international cooperation, and focus their attention and resources inward. They thus diminish opportunities for mutually beneficial economic exchange and forego the benefits of globalization, which can worsen both internal and global economic turmoil.
In Figure 9c, we indicate that, in the decades ahead, less international cooperation will perhaps fatally weaken international action to slow climate change. More frequent and severe extreme weather events will then trigger flows of migrants toward richer countries (Lustgarten, Reference Lustgarten2020; Xu et al., Reference Xu, Kohler, Lenton, Svenning and Scheffer2020), an influx that is likely to increase support for chauvinistic and isolationist ideologies in receiving societies. The resulting exacerbation of economic turmoil could ultimately propel out-migration from these countries.
Finally, Figure 9d shows that the chauvinistic reaction to mass migration is likely to precipitate violence against those seeking refuge and those deemed too sympathetic toward outsiders. Meanwhile, extreme weather events could worsen intercommunal tensions, trigger state collapse and civil war, and increase the probability of international conflicts over scarce resources, including water and food. Civil violence and interstate war tend to deepen nationalism while generating new waves of refugees and exacerbating economic turmoil. These pernicious feedbacks are certainly not inevitable; but if they were to take hold they would escalate all of the problems depicted in Figure 9 in a catastrophic spiral.
4.6 Mapping dominos and loops
The two mapping techniques illustrated above – one focusing on domino effects and the other on inter-systemic feedback loops – complement each other and together enable a distinctly network-based approach to crisis analysis. They help researchers identify those nodes (stresses, triggers, or crises) that are most influential – that affect many other nodes in the network – and those that are most vulnerable – that are most affected by other nodes (Low, Reference Low2021).Footnote xii In Figure 8, inflation is particularly vulnerable in this sense, while the pandemic is highly influential. Many of the feedback loops presented in Figure 9 travel through the nationalism and anti-globalization node, suggesting that people's ideological reactions to change will play a highly influential role in shaping future outcomes.
Network maps of global crises are not entirely new, of course. The World Economic Forum's (WEF) annual Global Risk Report has included similar diagrams since 2007 (WEF, 2007, p. 13). Most recently, the 2023 report (WEF, 2023) identifies ‘state collapse’, ‘erosion of social cohesion’, ‘collapse of a systemically important supply chain’, ‘interstate conflict’, and ‘cost-of-living crisis’ as the most influential global risks (i.e. those most connected to other global risks).
Although WEF's diagrams provide useful insights into the architecture of the current global polycrisis, the Forum acknowledges their limitations. For one thing, the diagrams' links depict only positive correlations between risks – that is, about the likelihood that certain risks will appear together – not actual causal connections between them.Footnote xiii Neither do the diagrams convey information about negative correlations, where the occurrence of certain risks diminishes the likelihood of others (WEF, 2008, p. 25). Also, network maps like those presented by the WEF provide a static snapshot of risk connections at a particular moment; they do not reveal how risks and their connections change over time as crises occur and activate other risks. Finally, the data underlying the WEF diagrams are derived from surveys of leaders and experts in the business community. But when scientists were asked to appraise the same global risks, they generally judged them to be more likely and harmful than did the WEF's respondents (Future Earth, 2020; Future Earth et al., 2021; Garschagen et al., Reference Garschagen, Wood, Garard, Ivanova and Luers2020).
To address these challenges, the emerging polycrisis research program should prioritize methodological innovation that uses valid and reliable measures of key variables to identify the actual casual mechanisms linking stresses, triggers, and crises.
5. Conclusion: a polycrisis research program and lessons for policy
We argue that the world is experiencing a worsening polycrisis and propose a conceptual framework for understanding how crises (and their precursor stresses and triggers) become entangled across global systems. This framework will help researchers identify and study the causal mechanisms that produce crisis amplification, acceleration, and synchronization.
We thus place the polycrisis concept at the center of an urgent new research program. This program can draw on theories and methods in other fields to explain the dynamics of crisis interaction. Complexity science provides theories explaining critical transitions (Scheffer, Reference Scheffer2009), path dependence (Pierson, Reference Pierson2004), stability landscapes (Folke et al., Reference Folke, Carpenter, Walker, Scheffer, Chapin and Rockström2010; Walker et al., Reference Walker, Holling, Carpenter and Kinzig2004), and the underlying sources of complexity (Arthur, Reference Arthur1993). Network science elucidates the structure of connectivity within global systems, including the interactions between networks (Buldyrev et al., Reference Buldyrev, Parshani, Paul, Stanley and Havlin2010; Gao et al., Reference Gao, Liu, Li and Havlin2015). And process tracing (a method of historical analysis in the social sciences) allows researchers to discern causal mechanisms in situations where controlled-case comparisons are impossible (George & Bennett, Reference George and Bennett2005, pp. 205–232), as when observed crisis interactions are historically unprecedented.
Targeted empirical research investigating specific crisis interactions can guide policymakers and other actors seeking to navigate the polycrisis. Our analysis points to three broad policy implications.
Focus on crisis interactions, not isolated crises: Governments tend to focus on individual and immediate threats, which often renders their management of systemic risks ineffective (ISC et al., 2022, p. 8). Because today's crises are causally entangled, they can be neither fully understood nor addressed in isolation from one another. A comprehensive approach is necessary – an ‘integrated assessment’ of the full range of interlinked crises involved – especially when policies that address one crisis might worsen or undermine efforts to resolve others (Baum & Barrett, Reference Baum, Barrett and Garrick2018).
Address system architecture, not just events: The polycrisis concept also highlights the role of densely interconnected global systems as the conduits that transmit the causes and effects of cascading crises. Policymakers must work to change system structures that generate such hazards. They can, for instance, strengthen negative (dampening) feedbacks that counteract pernicious positive feedbacks. In some cases, they might reduce connectivity or create buffers (or firebreaks) at sites of systemic vulnerability. Recent efforts to more heavily regulate financial institutions designated as systemically important show that governments and international agencies are starting to internalize this principle. But much more can be done to protect technological infrastructure (by increasing the resilience of vital communications systems to electrical grid and satellite failures, for instance), strengthen food systems (by buffering against the risk of simultaneous breadbasket failures (Gaupp et al., Reference Gaupp, Hall, Hochrainer-Stigler and Dadson2020)), and reduce pandemics arising from zoonotic spillover (by limiting wet markets, bushmeat consumption, and the illegal wildlife trade). Broadly speaking, policymakers should consider resilience alongside efficiency when evaluating policy outcomes, which means encouraging policy diversity, experimentation, and redundancy – all elements of adaptive management.
Exploit high-leverage intervention points: Many of the same features of complex systems that create polycrises also provide opportunities for systemic transformations toward more desirable futures. When systems are prone to non-linearities, positive feedbacks, and critical transitions, a relatively small action may have a profound effect, if it is tailored to the system's features (Lenton et al., Reference Lenton, Benson, Smith, Ewer, Lanel, Petykowski, Powell, Abrams, Blomsma and Sharpe2022; Meadows & Wright, Reference Meadows and Wright2008, pp. 145–165; Otto et al., Reference Otto, Donges, Cremades, Bhowmik, Hewitt, Lucht, Rockström, Allerberger, McCaffrey, Doe, Lenferna, Morán, van Vuuren and Schellnhuber2020). Network-based polycrisis visualization and analysis can help identify such intervention points.
The polycrisis concept – if effectively grounded in a scientific research program focused on practical steps to improve policy outcomes – can help us better address the world's interlinked crises. It can inform strategies to prevent the amplification, acceleration, and synchronization of crises and to respond when polycrises occur. But this research program needs to start now. ‘Business as usual’, says United Nations Secretary-General António Guterres, ‘could result in breakdown of the global order, into a world of perpetual crisis and winner-takes-all’ (Guterres, Reference Guterres2021).
Acknowledgements
None.
Author contributions
M. L. and T. H.-D. developed the polycrisis conceptual framework and wrote the text with input from all co-authors and editing by S. J.
Financial support
M. L. is supported by the V. Kann Rasmussen Foundation and an Omega Resilience Award. J. R. and J. F. D. acknowledge support from the European Research Council Advanced Grant project ERA (Earth Resilience in the Anthropocene, ERC-2016-ADG-743080). J. F. D. is grateful for financial support by the project PIK_Change (grant 01LS2001A) funded by the German Federal Ministry for Education and Research (BMBF).
Competing interest
Michael Lawrence, Thomas Homer-Dixon, Scott Janzwood, Johan Rockström, Ortwin Renn, and Jonathan F. Donges declare no conflict of interest.
Research transparency and reproducibility
Not applicable.