The fragility of interdependence: How globalisation turned complexity into systemic risk
· 19 min read
This is article 3 in The Complexity Paradox series. Here is article 2.
Globalisation transformed the scale at which modern societies could organise economic life. Production moved beyond national boundaries, capital could cross continents in seconds, and companies learned to source components from specialised suppliers scattered across multiple regions. Consumers gained access to products, technologies, and services that would once have been limited by geography. Shipping networks connected ports into global trade corridors, financial markets linked savings in one country to investment in another, and digital infrastructure compressed distance even further by allowing information, capital, coordination, and increasingly services themselves to move almost instantaneously. The result was not simply more trade. It was the construction of a global system in which production, finance, logistics, energy, technology, and information became deeply interwoven.
The gains were enormous. Globalisation allowed firms to specialise, reduce costs, reach new markets, and combine capabilities that no single country or company could efficiently reproduce alone. Countries could integrate into global value chains without having to build every stage of production domestically. Technologies diffused faster, international investment expanded, logistics networks became more sophisticated, and agricultural markets became more connected. Hundreds of millions of people entered new forms of economic activity as production and investment expanded across regions that had previously been more isolated from global markets. Yet the same architecture that enabled this prosperity also transformed the nature of risk, because globalisation did not merely connect markets. It connected vulnerabilities.
A manufacturing disruption in one country could halt production in another. A banking crisis could spread through financial markets within hours. A conflict affecting a major energy producer could alter costs across multiple continents, while a drought in one agricultural region could influence food prices far beyond the area experiencing the actual climate shock. A cyberattack against a major service provider could disrupt firms in dozens of countries, and a blockage at a strategically important shipping corridor could delay supply chains around the world. The international system became more capable because it became more interconnected, but that same interconnection also created pathways through which disruption could travel.
This is the fragility of interdependence. The problem is not interdependence itself. Interdependence can be an extraordinary source of resilience because it allows societies to draw on resources, capabilities, knowledge, markets, and institutions beyond their own borders. A country experiencing crop failure can import food, a region facing an energy shortage can draw on external supply, and a company affected by local disruption may be able to shift production elsewhere. International finance can mobilise capital at a scale that would be impossible through domestic savings alone. The danger emerges when interdependence becomes concentrated, opaque, tightly coupled, or difficult to substitute. Under those conditions, systems that appear efficient during periods of stability can become channels for systemic failure during periods of stress.
Globalisation therefore created a paradox similar to the one explored throughout this series. The systems that reduced the limitations of geography also reduced the containment of disruption. The same structures that allowed prosperity to spread more quickly also allowed shocks to spread farther.
Ulrich Beck's concept of the "risk society" offers a useful way to understand this transformation. Beck argued that advanced modern societies increasingly produce risks through the very processes of modernisation that generate prosperity. Industrial development, technological innovation, scientific advancement, and global integration solve problems while simultaneously producing new forms of uncertainty that are difficult to contain within traditional geographic, institutional, or political boundaries.
This insight is particularly relevant to globalisation because many contemporary risks no longer belong neatly to individual countries. Climate change crosses borders. Financial crises move through international markets. Cyber risks travel through global digital infrastructure. Supply disruptions propagate across global value chains. Pandemics move through international transportation networks. Environmental contamination can affect regions far from its source, while strategic competition over technology, minerals, energy, and data increasingly affects economic systems far beyond the states directly involved.
Traditional governance, however, was built largely around territorial jurisdictions, while many of the systems generating contemporary risk operate through networks. Governments regulate within borders, but finance, supply chains, data, energy markets, logistics, and digital platforms often operate across them. Companies may have legal headquarters in one country, suppliers in several others, customers across dozens more, cloud services hosted elsewhere, and financial exposure spread through institutions operating globally. The system functions internationally while accountability remains fragmented among national jurisdictions.
This mismatch between networked systems and territorially bounded governance makes it difficult to identify where responsibility begins and ends. A disruption may originate in one country, propagate through corporate networks, pass through financial markets, and ultimately affect consumers in places with no direct connection to the original event. This is why systemic risk differs from ordinary risk. Ordinary risk can often be understood as exposure to a particular event, but systemic risk concerns the possibility that disruption will move through relationships within the system itself. The danger is not simply the initial shock. It is the structure through which that shock is transmitted.
Globalisation expanded that structure enormously.
Charles Perrow's work on "normal accidents" provides another important perspective. Perrow examined systems that are both highly complex and tightly coupled. In such systems, interactions between components may be difficult to anticipate, while failures in one part can quickly affect others before operators have time to understand or contain what is happening. The concept originated in the study of technological systems, but its logic applies surprisingly well to the global economy.
Modern global production systems are extraordinarily complex. A finished product may depend on hundreds or thousands of suppliers distributed across multiple tiers, and many of those suppliers themselves depend on other suppliers whose identities may not be visible to the company selling the final product. Logistics providers depend upon ports, fuel, communications systems, customs authorities, software platforms, insurance markets, and shipping capacity. Financial institutions depend upon counterparties, payment networks, clearing systems, digital infrastructure, and confidence. What appears from the outside to be a single product or transaction is often the visible endpoint of a much larger network.
These systems are also increasingly tightly coupled. Firms minimise inventory to reduce costs, production schedules are synchronised across continents, financing depends on continuous liquidity, transportation networks operate within narrow timing windows, and digital systems process transactions in real time. Under normal conditions, this creates extraordinary performance. Under stress, however, it can reduce the time available for adaptation. A company with months of inventory has time to respond to supply disruption, while a company with only days of inventory may not. A financial institution with significant liquidity reserves has more room to absorb shocks than one operating with minimal buffers, just as a power system with excess capacity can manage unexpected demand more easily than one operating close to its limits.
The issue is not that efficiency is undesirable. Efficiency is one of the reasons global economic integration has produced such enormous gains. The problem arises when efficiency is achieved by eliminating redundancy, buffers, and alternative pathways to the point that the system becomes brittle. Complexity makes failures difficult to anticipate, and tight coupling makes them difficult to contain. Together, they create conditions in which relatively small disruptions can produce consequences far larger than the original event.
Global value chains are one of the defining features of modern economic organisation. They allow different stages of production to occur where comparative advantages are strongest. Research may take place in one country, design in another, raw materials may come from several others, components may be manufactured in specialised industrial clusters, assembly may occur elsewhere, and final products may be distributed globally. This fragmentation of production has generated remarkable efficiency and specialisation, but it has also created a hidden geography of dependency.
The final location of a company tells us relatively little about where its true risks reside. A firm headquartered in Europe or North America may depend on mineral extraction in Africa, semiconductor fabrication in East Asia, software development in India, logistics hubs in the Middle East, cloud infrastructure operated by a small number of global providers, and shipping routes passing through narrow maritime corridors. This means corporate risk increasingly exists outside the boundaries of the corporation itself.
The vulnerability may be three or four tiers down the supply chain. A company may believe it has diversified suppliers because it buys from several firms, only to discover that those firms all rely on the same upstream producer. Different suppliers may use the same port, the same energy source, the same shipping route, the same cloud provider, or the same critical material. What appears diversified at one level may be concentrated at another, and that concentration can remain hidden until disruption reveals it.
The challenge is compounded by the fact that global supply chains evolved primarily around cost, reliability, scale, and speed rather than transparency. Companies often understand their direct suppliers reasonably well but have limited visibility beyond the first tier. The further one moves upstream, the more difficult the network becomes to map, and this matters because shocks frequently emerge from precisely those less-visible layers. A shortage of a relatively inexpensive component can halt production of a much more valuable product, while a disruption at a specialised upstream supplier can affect dozens of downstream industries. A problem at a single logistics node can create delays across multiple supply chains.
The economic importance of a component is therefore not determined simply by its price. It is determined by whether the system can function without it.
Few management innovations illustrate the trade-off between efficiency and resilience as clearly as just-in-time production. The logic is compelling: inventory costs money, products sitting in warehouses tie up capital, require space, create insurance and management costs, and risk becoming obsolete. By synchronising production more closely with demand, firms can reduce inventory, improve cash flow, and operate more efficiently. Across thousands of companies, these gains become enormous.
But inventory is not only a cost. It is also a buffer.
When production is organised around highly synchronised flows, disruption can quickly move downstream. A delayed shipment, factory shutdown, transportation interruption, or supplier failure can begin affecting production almost immediately because there is little inventory available to absorb the shock. This does not mean just-in-time production is inherently fragile. In stable environments with diversified suppliers, strong logistics, and good visibility, it can function extremely well. The vulnerability emerges when the assumptions supporting that efficiency no longer hold.
This distinction matters because debates about resilience can become overly simplistic. The answer is not necessarily to abandon just-in-time systems and accumulate large inventories everywhere, because doing so would create enormous costs and inefficiencies. The more useful question is where buffers matter most. Not every input deserves strategic redundancy. Some are easily substitutable, while others are critical. Some suppliers can be replaced quickly, while others may require years of investment before alternatives become available.
Resilience therefore requires differentiation. The system must know which dependencies are ordinary and which are strategic. Without that understanding, organisations can optimise themselves into fragility.
Financial globalisation illustrates perhaps the purest form of systemic interdependence because financial systems are built largely on relationships of obligation, confidence, and liquidity. Banks lend to one another, investors hold assets across multiple markets, companies borrow internationally, currencies move through global trading systems, and derivatives link institutions through contractual obligations. Asset prices influence collateral values, which influence lending capacity, which in turn influence economic activity.
Finance therefore contains powerful transmission mechanisms. A decline in one market can force investors to sell assets elsewhere. Losses at one institution can create uncertainty about counterparties. Uncertainty can produce liquidity hoarding. Falling asset prices can weaken balance sheets, while weaker balance sheets can reduce lending. Reduced lending can then affect businesses and households far removed from the financial institutions where the initial problem began.
In highly connected financial systems, confidence itself becomes a form of infrastructure. When confidence disappears, relationships that functioned smoothly can suddenly freeze. This is why financial crises can propagate with such speed. Capital moves far more quickly than physical goods, and expectations can change almost instantaneously. A problem does not need to spread physically because it can spread through information, fear, balance sheets, and market behaviour.
Network theory helps explain why this matters. Highly interconnected networks can be robust to many small failures because participants have multiple relationships and alternative pathways, but networks can also become vulnerable when certain institutions function as major hubs. If a peripheral node fails, the system may absorb it. If a highly connected hub fails, the consequences can be much larger.
The same logic applies beyond finance. Ports, cloud providers, semiconductor manufacturers, telecommunications networks, payment systems, and energy infrastructure can all function as hubs within larger systems. The structure of the network therefore matters as much as the number of participants within it.
One of the ironies of globalisation is that an economy spanning the planet still depends on a relatively small number of physical corridors. Maritime trade is global, but ships must often pass through narrow straits and canals. Energy flows through pipelines, terminals, and shipping lanes. Digital communications depend on terrestrial networks, satellites, and subsea cables. Air transportation relies on major hubs, while global logistics depends heavily on large ports capable of handling immense volumes of cargo.
Globalisation reduced the economic importance of distance, but it did not eliminate geography. In some cases, it increased the importance of particular geographic nodes. Shipping chokepoints are a clear example. A narrow waterway may carry trade between vast economic regions, and when that route is disrupted, ships may be forced to travel thousands of additional kilometres, increasing transportation time, fuel consumption, insurance costs, and supply uncertainty.
The same logic applies to ports. A highly efficient port can become indispensable precisely because so much activity is concentrated there. Scale creates efficiency, but scale can also create systemic importance. Digital infrastructure contains similar chokepoints, although they are less visible. Subsea cables carry enormous volumes of international data, cloud computing is concentrated among a relatively small number of large providers, and satellite infrastructure supports communications, navigation, logistics, agriculture, and defence. A failure in one of these systems may affect activities that appear unrelated because they rely on shared infrastructure beneath the surface.
The lesson is not that chokepoints can be eliminated. Many exist because geography, economics, and technology naturally create concentration. The challenge is to understand where they exist and what alternatives are available when they fail.
The energy transition is creating another layer of interdependence around critical minerals. Decarbonisation requires enormous expansion of technologies such as batteries, electric vehicles, renewable energy systems, transmission infrastructure, and advanced electronics. These technologies depend on minerals whose extraction, processing, and refining are often geographically concentrated.
This creates an important development paradox of its own. Reducing dependence on fossil fuels can create new dependencies on mineral supply chains. The environmental logic of the transition remains strong, but its resilience depends upon understanding these new relationships. A country may reduce dependence on imported oil while increasing dependence on imported batteries, processed minerals, specialised components, or manufacturing capacity.
Again, the issue is not dependency itself. International exchange allows resources to be used far more efficiently than attempting to reproduce every capability domestically. The risk emerges when supply is highly concentrated, substitution is difficult, production requires long lead times, and political or economic disruption can limit access. Critical minerals therefore reveal a broader principle: transitions in complex systems rarely eliminate dependency. They redistribute it.
Energy security in the twentieth century was shaped heavily by oil and gas. Energy security in the twenty-first century will increasingly involve minerals, grids, batteries, software, manufacturing capacity, transmission infrastructure, and digital control systems. The map of dependency changes, but dependency remains.
Food and energy markets demonstrate how systemic risks can move between sectors. Agriculture depends on energy for fertiliser production, irrigation, machinery, processing, refrigeration, and transportation. Energy prices therefore influence food prices, while financial conditions influence farmers' ability to purchase inputs. Climate events influence crop yields, trade policies influence market availability, and currency movements affect import costs. A shock originating in one system can therefore travel through several others before reaching consumers.
The same is true in energy markets. Industrial production depends on energy, transportation depends on energy, data centres depend on energy, water systems often depend on energy, and agricultural production depends on energy. This means energy disruption can become industrial disruption, food disruption, communications disruption, or inflation.
Systemic risk is often difficult to understand because the final effect can appear far removed from the original cause. A drought can become an electricity problem if hydropower production falls. An electricity shortage can become an industrial problem, an industrial slowdown can become an employment problem, and an employment shock can become a fiscal problem. Fiscal stress can then reduce public investment and social protection. This is how complex adaptive systems behave: causes and consequences do not remain within sectoral boundaries.
The international system contains countless feedback loops of this kind, and globalisation has increased the number of connections through which those effects can travel.
If shipping networks created the physical infrastructure of globalisation, digital networks increasingly form its nervous system. Financial markets, logistics, communications, government administration, industrial systems, aviation, energy networks, hospitals, supply chains, and businesses now depend on digital infrastructure. Cloud computing allows organisations to access extraordinary computing capacity without building it themselves. Software platforms coordinate transactions across borders, satellites support navigation and communications, and artificial intelligence is beginning to influence decision-making across an expanding range of sectors.
These systems create capabilities that would have been unimaginable only a generation ago, but they also create shared dependencies. A company may believe it operates independently while relying on the same cloud provider as thousands of other organisations. Governments may depend on commercial software systems used across multiple jurisdictions, global logistics may rely on a limited number of digital platforms, and navigation systems may depend on shared satellite infrastructure.
When infrastructure becomes shared, disruption can become systemic. Cyber risk is therefore not simply about individual organisations protecting themselves from hackers. It is increasingly about the resilience of the digital architecture on which entire sectors depend. A cyberattack against one company is a corporate problem, but a cyberattack against infrastructure used by thousands of companies becomes a systemic problem.
The distinction matters because responsibility for resilience becomes more difficult to allocate. Individual organisations can improve cybersecurity, but they cannot fully control risks embedded in infrastructure outside their boundaries. This is another characteristic of interdependence: resilience cannot always be produced individually. It must sometimes be produced collectively.
The fragility of global systems should not be confused with an argument for deglobalisation. Interdependence creates vulnerabilities, but it also creates resilience. A country facing local crop failure is safer when it can import food than when it depends entirely on domestic production. A company facing disruption at one supplier is safer if alternative suppliers exist in other regions. A government responding to disaster benefits from international financing, expertise, technology, and humanitarian networks. Energy interconnections can allow neighbouring systems to support one another during shortages.
The critical issue is therefore the structure of interdependence. Diverse networks can be resilient, while concentrated networks can be fragile. Transparent networks can be managed more effectively than opaque ones. Loosely coupled systems provide more time for adaptation, while tightly coupled systems transmit failure more rapidly. Redundant systems can reroute around disruption, while systems dependent on single points of failure cannot.
The distinction between interdependence and fragile interdependence may therefore be one of the most important distinctions in contemporary risk management. The policy debate often becomes trapped between two extremes. One argues that deeper globalisation automatically produces prosperity and should therefore continue with minimal interruption. The other argues that dependence itself is dangerous and that resilience requires greater self-sufficiency.
Both positions miss the structural issue. No modern economy can realistically become independent of global systems without sacrificing enormous economic and technological benefits. At the same time, assuming that markets will always provide substitutes quickly enough during disruption ignores the realities of concentration, infrastructure, geography, and production lead times. The challenge is not independence. It is resilient interdependence.
A more resilient global system would not necessarily be less connected. In some respects, it might be more connected, but connected differently. Resilience can come from diversification across regions, suppliers, technologies, and transportation routes. It can come from better visibility into supply chains, strategic inventories for critical inputs, interoperable systems that allow substitution when one provider fails, and distributed production capacity that prevents excessive concentration.
It can also come from better governance. Governments need greater visibility into strategic dependencies. Companies need to understand risks beyond their direct suppliers. Financial regulators need to understand network effects rather than focusing only on individual institutions. Cybersecurity policy must address shared infrastructure. Energy planning must consider how transitions create new forms of dependency. Development policy must consider whether integration into global value chains creates durable capability or simply new exposure.
This requires a different way of measuring success. For decades, efficiency was often measured through cost reduction, inventory reduction, increased trade, greater specialisation, and faster flows of capital and goods. Resilience requires additional questions about how many alternative suppliers exist, how quickly production can shift, where single points of failure are located, which nodes are systemically important, how much time the system has to adapt after disruption, and which dependencies are visible versus hidden.
These questions do not replace economic analysis. They deepen it.
Globalisation succeeded in part because it allowed societies to overcome the limitations of geography. Resources, expertise, production, finance, information, and technology could be combined across borders in ways that dramatically expanded human capability. But interconnection also changed the geography of risk because disruption could now move farther and faster through the same networks that supported prosperity.
The same networks that move goods can move scarcity, the same financial systems that move capital can move contagion, and the same digital infrastructure that moves information can move cyber disruption. Energy markets that allow countries to access global supply also transmit price shocks, while food markets that protect societies from local shortages can also transmit volatility across regions.
This is not evidence that globalisation failed. It is evidence that globalisation matured into a complex system whose risks must be governed differently from those of the world that preceded it. The international system is not fragile because it is connected. It becomes fragile when those connections are concentrated, tightly coupled, poorly understood, difficult to substitute, and governed without sufficient visibility into their systemic consequences.
Interdependence can distribute risk or concentrate it. It can create redundancy or dependence, provide alternatives or create chokepoints, and absorb shocks or amplify them. The outcome depends upon the architecture of the network. The challenge for the next phase of globalisation is therefore not to choose between integration and isolation, but to redesign interdependence so that prosperity does not require fragility.
That means preserving the extraordinary benefits of specialisation and exchange while building more diversity, visibility, redundancy, adaptive capacity, and institutional understanding into the systems that support them. The twentieth century built an international economy optimised to move goods, capital, information, and energy with unprecedented efficiency. The twenty-first century must learn how to ensure those systems can continue functioning when the unexpected happens.
In a deeply interconnected world, resilience is no longer simply the ability of an individual country, company, or institution to withstand disruption. It is the ability of the network itself to absorb failure without allowing one shock to become everyone's crisis.
illuminem Voices is a democratic space presenting the opinions of leading Sustainability Thought Leaders, their views do not necessarily represent those of illuminem.
The world needs sustainability knowledge. At illuminem, no interest group or shareholder can influence our work. Thank you for supporting our mission to make high-quality and independent sustainability information free for all. Every contribution helps. Thank you for donating today.
Steven W. Pearce
Public Governance · Sustainability Law
Kasper Benjamin Reimer Bjørkskov
AI · Ethical Governance
illuminem briefings
AI · Public Governance
The Guardian
AI · Public Governance
Politico
Public Governance · Climate Change
Carbon Herald
Carbon Market · Public Governance