This week's in-depth discussion – System collapse with Michael Sahlin and Lars-Erik Lundin
Here is a more detailed text richly illustrated with images and media that highlights a central topic for the Academy's project activities from several different perspectives.
The article is intended to provide a basis for conversation and debate, both within academia and in the wider public.
The worst thing that could reasonably happen is a nuclear war. This is a widespread belief, and the logic of deterrence makes it likely that this is precisely why such a war has not occurred so far.
But that also brings with it another problem.
Nuclear deterrence seems to create a stability that — paradoxically — can make another type of catastrophe possible: that the international system that since 1945 has limited great-power behavior, coordinated global collective goods, and protected the room for action of small states is eroded piece by piece, at a pace that rarely triggers a collective response. Few individual events are perceived as a system shift. The consequences can still be similar.
This text based on a conversation between Michael Sahlin and Lars-Erik Lundin and on extensive background material is an attempt to clarify what such erosion could mean, why the US's role change appears to be one of its most critical driving forces, and why Sweden, in our assessment, risks preparing for a partly wrong threat picture.
Not a disaster — slightly slower
There is a common simplification of what “global system collapse” would mean. The term is easily associated with a third world war or a single event that destroys everything. But a more reasonable reading is that global system collapses rarely resemble single events — they are more like processes.
Operationally, the concept in this text can be understood as a process in which the international regimes and institutions that regulate great power behavior and protect the room for action of small states gradually erode to a level where they fulfill their functions increasingly poorly and in some scenarios may fail when they are needed most — without the collapse of nuclear deterrence and without a single kinetic event initiating the process.
In that case, it would be a gradual erosion — the kind that is often the hardest to stop, precisely because no single decision triggers it.
The World Economic Forum, in its Global Risks Report 2026, assessed the likelihood of serious disruptions in global institutions as high over a five-year period. SIPRI's 2025 yearbook points in the same direction, describing a gradual erosion of the arms control regimes built up since the 1970s: New START has expired, the INF Treaty has ceased to exist, and China's arsenal is estimated at over 750 warheads and is expected to grow.
This is less a forecast than a situation description of the direction — but how far the process will go is not given.
| FACTBOX
System collapse as sequential link breaking Global system collapse does not have to occur through a single catastrophic breach, but can conceivably occur through a gradual erosion where subsystem after subsystem fails without a collective response being triggered. In a flow security perspective, this could mean that each cable incident that is normalized without institutional consequence tends to lower the threshold cost of the next attack. Three cascading mechanisms can conceivably reinforce the process: resource competition (crisis management in one system can drain resources from neighboring systems), legitimacy contagion (failure in one domain can undermine confidence in protection in the next) and norm transfer (repeated norm violations without a response can normalize the next violation). Source: Wedin, Lars, “Flow Security — a Concept for Our Time,” Proceedings and Journals of the Royal Swedish Academy of Military Sciences, January 2025; Farrell & Newman, “Weaponized Interdependence,” International Security 44(1), 2019. |
In the conversation with Michael Sahlin, the concept was formulated as follows:
| System collapse — the concept and mechanism
Michael Sahlin defines the analytical framework from which Project SV-AR 2.0 works: not a single war but a gradual erosion of global rule systems. STARTING POINT Globalization has created a tightly interconnected macrosystem. Nuclear deterrence creates stability in one dimension but at the same time enables a long-term race under the nuclear ceiling — a gray area where subsystem after subsystem can be eroded without a single breach triggering a collective response. KEY POINTS • Subsystem logic. The world consists of a number of subsystems — nuclear arms control, trade, climate, cyber governance, transatlantic security — with connections between the parts and the whole. • Gradual or cascading. Collapse can occur slowly (trickle down) or suddenly (cascade). Both forms are real; the planning horizon must cover both. • The paradox. The nuclear threat is the worst possible thing, but the fear of it creates space for the kind of hybrid competition and institutional erosion that we call systemic collapse. CONCLUSION System collapse is not an apocalypse — it is a process. It is gradual enough for political systems to underestimate it, but rapid enough that preparedness gaps do not have time to close in time. Source: Conversation between Lars-Erik Lundin and Michael Sahlin, 2026-05-29, 33:00–37:52 |
The nuclear weapons paradox
There is a possible connection between nuclear deterrence and systemic collapse that is rarely discussed openly: they may, to some extent, be prerequisites for each other. Nuclear deterrence appears to have worked so far. There is currently little evidence that great powers are prepared to escalate to strategic nuclear use. But this very thing seems to create a space below the nuclear ceiling — a gray area where actors can act aggressively, erode institutions, break rules, and project power without triggering the collective response that a kinetic attack would likely require. The reasoning can be seen as a broadening of the classic stability–instability paradox: strategic nuclear stability reduces the risk of direct great-power war, but can at the same time increase the scope for aggressive sub-threshold actions and institutional erosion.
Hybrid operations against the Baltic Sea's underwater infrastructure, the Houthi movement's disruption of trade via the Red Sea (according to some estimates around twelve percent of world trade), and Russian influence on Swedish and Finnish opinion can be seen as examples of actions that are destabilizing but not necessarily system-triggering. There is much evidence to suggest that they are designed to be just below the threshold for collective response.
Our assessment is that nuclear deterrence does not in itself constitute a secure protection against systemic collapse, and that in some sense it may even contribute to making it possible. As long as the strategic nuclear threat persists, great powers seem to be able to engage in sub-strategic erosion without facing proportionate countermeasures.
It is a paradox that any reasoning about system collapse reasonably needs to address.
| FACTBOX
The Baltic Sea bottom — the binding weak link According to common estimates, a very large proportion — often around 95 percent — of Europe’s international internet traffic and a large part of the region’s energy flows pass through optical fibre cables and pipelines, many of which run along the bottom of the Baltic Sea. The repair time in the event of damage is often given as between one and four weeks. The series of incidents in 2022–2024 is consistent with a worrying pattern: Nord Stream (September 2022), Baltic Connector (October 2023), Estlink-2 (December 2024) and the Arelion/Cinia cable and BCS East-1 (November 2024) — the last case of which has been assessed by the EU as likely coordinated hostile actions. There is much evidence to suggest that the Baltic Sea bottom domain is among the most under-monitored and poorly protected links in the flow system of northern Europe. Source: EU Action Plan for Cable Security, 2024; ENISA Threat Landscape Report 2025; EU Commission statements on Estlink-2 and Arelion/Cinia, November–December 2024. |
| FACTBOX
Cascade — when one flow fails, all are affected The flow domains seem to be interconnected in a way that political analysis rarely fully reflects. When the Houthi movement attacked merchant ships in the Red Sea in 2023–2024, Suez Canal traffic was reported to fall by about half during periods and freight costs on some routes to rise sharply, in some estimates by several hundred percent — and a sinking ship during the conflict is also said to have damaged data cables along the same bottom profile. A shipping disruption can thus affect trade flows and data flows more or less simultaneously. Something similar seems to apply to the bottom geography of the Baltic Sea, where cable routes for data traffic partly run parallel to energy pipelines. Sabotage against one type of infrastructure can thus create exposure for neighboring ones as well. Source: IMF Working Paper, “Shipping Cost Shocks and Global Supply Chains,” 2024; EU Action Plan for Cable Security, 2024; Farrell & Newman, “Weaponized Interdependence,” International Security 44(1), 2019. |
USA: from load-bearing beam to erosion driver
The rules-based international system was essentially built under American leadership after 1945. Bretton Woods, the UN Charter, NATO, NPT, GATT and the WTO can largely be understood as American-backed institutional projects, driven by a hegemon that seems to have counted on a stable global rules system serving its own interests.
That calculation seems to have held up for decades. There is reason to believe that it no longer does so to the same extent.
The US under the current administration appears not only as an actor that in some respects is reneging on previous commitments, but also as an actor that in several policy areas appears to be contributing to the erosion of the systems it once guaranteed.
There is much evidence to suggest that this is not just a temporary anomaly, but at least partly reflects a longer-term shift in how the US national interest is defined.
The experience of three decades of European security policy leads us to the assessment that Europe does not yet seem to have fully grasped what it would mean if the US were no longer the guarantor of the system to the same extent as before. Europe is used to demanding more from the US and criticizing US actions. Planning for a Europe with a significantly weaker American anchor is something else.
In the conversation, Sahlin describes the US's changing role as the single most destabilizing factor:
| USA — from load-bearing beam to erosion driver
Sahlin identifies the US's changing role as the single most destabilizing factor in the global system in 2026. STARTING POINT The international systems — the UN, WTO, NPT, NATO, the Paris Agreement — are built around claims to legitimacy and underlying guarantors of power. The role of hegemon has historically been carried by the United States. KEY POINTS • Erosion of the rules of the game. Player after player is putting self-interest before obedience to established rules of the game — Netanyahu, Russia, and now the United States under the current leadership. • Sahlin emphasizes the distinction: it is not just that the US no longer protects the system — the US is actively helping to erode it. • The house of cards effect. Nuclear deterrence may continue to work, but that does not rule out the collapse of the rest of the system — a house of cards with an intact foundation but collapsing walls. CONCLUSION System collapse does not have to mean nuclear war. It is enough for the US to actively erode the systems that were once built to prevent war. Source: Conversation between Lars-Erik Lundin and Michael Sahlin, 2026-05-29, 43:44–45:11 |
Corruption adds a non-linear dimension to the same role change:
| Corruption as a tipping point mechanism
Sahlin highlights corruption's ability to overthrow otherwise predictable power structures as an underestimated element of systemic collapse. STARTING POINT Fareed Zakaria's analysis of the Trump administration from a corruption perspective leads Sahlin to a broader thesis: massive corruption under the guise of power creates tipping points that destroy otherwise analyzable lines of development. KEY POINTS • Black swans. Trump, Russia, Turkey, China — in each of these cases there is a tipping point possibility where the possession of power itself arouses enough irritation, suspicion, or anger to overturn the order. • The Orbán effect. Hungarian elections 2026: Magyars mobilized against corruption and Russian dependence — not against the EU. European far-right parties are moving against nationalism rather than global Trumpism. • Renationalization. The net result of these movements is a renationalization of European thinking — not necessarily anti-European, but suspicious of both Brussels, Moscow, and Washington. CONCLUSION Corruption is not just a political science problem — it is a system-destabilizing force with the ability to create non-linear events on a global scale. Source: Conversation between Lars-Erik Lundin and Michael Sahlin, 2026-05-29, 1:13:00–1:13:42 |
Weakest link
Against this background, historic investments are now being made in the European defense industry. Sweden is increasing its total defense budget to five percent of GDP. The EU's ReArm Europe initiative aims to mobilize 800 billion euros by 2030. NATO's new planning cycles address sustainability gaps identified after the experience in Ukraine.
It appears necessary. But it is hardly sufficient in itself.
Economist Chad Jones has argued in notable contributions that economies with weakest-link logic — where the aggregate outcome is largely determined by the least productive input — can behave significantly differently from classical production models. The asymmetry is striking: benefits seem to require the entire chain to be intact, while a single link breaking can be enough to cause harm.
Defense systems can largely be said to follow that logic. An investment of one hundred billion in military equipment risks providing very little actual capability if the decision-making chain is flawed.
Our assessment is that one of the most critical weak links in the Swedish and European system is likely not primarily resources, industry or even technology, but rather decision-making ability during a crisis.
Can the European Council coordinate during an acute scenario where several subsystems fail more or less simultaneously? Can Sweden, with the constitutional framework set by RF 12:2, among other things, make the decisions that may be required in a rapidly changing threat landscape? In our assessment, these questions lack an obvious "yes" — and therein lies a possible planning deficit.
An analogy: In 1972, the Swedish Institute for International Affairs conducted an interview survey with Swedish security policy elites — members of parliament, government representatives, analysts (Tom Hart and Lars-Erik Lundin). Out of eighty respondents, only one mentioned the word “energy.” The energy crisis came a year later.
What are we not mentioning today?
| FACTBOX
The Weakest Link Theory—What It Says Economist Chad Jones (Stanford) argues in a 2011 framework, later expanded to AI Economics 2026, that the overall output of a system can be largely determined by its weakest link rather than by the average. If a single function in a complex chain is missing, the value of the others can effectively be neutralized. In one of his calculations, Jones estimates that almost complete automation of half of all economic tasks would increase total output by only about 19 percent, as the remaining, hard-to-replace links then set the ceiling for the entire system. Source: Chad Jones & Christopher Tonetti, “AI and the Future of the US Economy,” NBER Working Paper 34779, March 2026; Jones, “Intermediate Goods, Weak Links, and Superstars,” American Economic Journal: Macroeconomics 3(2), 2011. |
| FACTBOX
Asymmetry — damage does not require full chain The weakest-link logic appears to be asymmetrical: benefits seem to require the entire chain to function, while damage can be caused by breaking a single node. Sabotage of a single submarine cable, pipeline or port can in principle paralyze a flow system whose other components are intact. It is partly this asymmetry that can make underwater infrastructure an attractive target: an attacker does not necessarily need to overwhelm the entire defense, but can target the binding weak link. Source: Jones, “Intermediate Goods, Weak Links, and Superstars,” AEJ: Macroeconomics 2011; EU Action Plan for Cable Security, European Commission, 2024. |
The reasoning about the weakest link was developed in the conversation with particular application to defense planning:
| Weakest link in defense planning
Lars-Erik Lundin applies Chad Jones's weak-link theory to Swedish and European defense planning: large investments can be rendered worthless by a single critical weak link. STARTING POINT Even though Sweden, the EU and NATO are investing enormous resources in defense and total defense, there is a risk that the system will nevertheless end up in a situation where nothing works — not due to a lack of resources, but because the weakest link is broken. KEY POINTS • Decision-making. The most central weak link at the Swedish level is the ability to make quick decisions. RF 12:2, the NSS vacuum and the capacity of the civilian leadership are concrete manifestations. • European level. Can the EU make decisions during an acute crisis? Can the European Council coordinate? These are not rhetorical questions but real planning assumptions. • Asymmetry. In weak-link logic: benefits require the entire chain to be intact, but damage requires only one broken link. Asymmetry argues for prioritizing the most critical links, not the broadest structure. CONCLUSION Defense planning must systematically identify and prioritize the weakest links — not the most visible or politically convenient. Source: Conversation between Lars-Erik Lundin and Michael Sahlin, 2026-05-29, 40:09–41:34 |
The conversation also pointed to a nearby vulnerability — that modern wars tend to be protracted:
| Modern wars are chronic wars
Michael Sahlin notes that the Ukraine war — longer than World War I — illustrates a structural pattern: modern wars do not end once they are allowed to break out. STARTING POINT The war in Ukraine has lasted longer than World War I. It is not an exception — it is a pattern. Wars that are allowed to erupt when the restraining mechanisms have eroded tend to be long wars. • The Missile Duel. Modern warfare is characterized by a continuous duel between missiles and missile defenses at all levels — from drones to nuclear weapons. Means drive countermeasures at an accelerating pace. • Simple, cheap weapons. Mass-produced drones represent a new type of asymmetry: cheap offensive systems versus expensive defensive ones. Europe is structurally underfunded in this dimension. • Eroded inhibiting mechanisms. The Ukraine and Iran wars illustrate how systemic collapse and long wars mirror each other: eroded institutions enable wars to break out; broken wars further erode institutions. CONCLUSION The planning assumption that a war in Europe would be short and limited lacks empirical support in 2026. Endurance is a strategic requirement, not a desire. Source: Conversation between Lars-Erik Lundin and Michael Sahlin, 2026-05-29, 53:37–55:03 |
What actually protects
There are hardly any simple measures. At the same time, uncertainty in itself does not justify inaction.
After this description of the situation, the reasoning here turns to what, in our opinion, can and should actually be done in Swedish and European planning.
Three areas appear to be particularly important to prioritize.
- Decision-making architecture during polycrisis. The EU’s institutional design appears to be largely adapted to sequential rather than simultaneous crises. The European Council has shown that it can act quickly — the coronavirus pandemic and the invasion of Ukraine are examples — but the structure for coordinating military, economic and informational matters during a compound crisis appears to be less developed. The NB8 cooperation (the eight Nordic-Baltic countries) may be one of the most readily available forums for regional polycrisis planning.
- Weakest link analysis in national planning. There are good reasons for Sweden to more systematically map its weakest links — not primarily the most obvious ones, but those that risk breaking silently. This could apply to decision-making levels for civil defense, dependence on foreign suppliers in the pharmaceutical sector, and seasonal variations in the energy system. There are already established methods (including ENISA's Critical Infrastructure Analysis and the EU's NIS2 Directive) that could be applied more broadly.
- Analytical capacity for systemic risks. Sweden currently appears to lack a body that continuously and systematically analyzes global systemic risks and their consequences for Swedish security — across the drainpipes and outside the secrecy framework of the intelligence services. The MSB's mandate does not appear to fully encompass economic security. The Defense Committee's report 2023 can be seen as a step in the right direction, but does not alone provide the ongoing analytical capacity that may be needed.
Three approaches, however, appear to be ineffective: essentially waiting for the US to return to its previous role, dealing with global systemic collapse as a specialist foreign policy problem, and allowing resource increases to replace strategic clarity.
| FACTBOX
Just-in-time as strategic vulnerability Modern society relies heavily on just-in-time logistics: goods are delivered close to the time they are needed, inventory is minimized, and the system is optimized for cost-efficiency under normal conditions. A typical grocery store is often said to have food for about two to three days, and Gotland needs to be replenished almost daily. A very large part of Sweden's imports — according to common data, over 90 percent — is transported by sea. This means that a cyberattack on a port, sabotage of a railway or a blockade of a sea route can, in unfavorable cases, disrupt supplies quickly — without a single soldier having crossed a border. Just-in-time thus appears not as a vulnerability among others, but as a design principle that can make the system particularly sensitive to precisely the type of disruption that gray zone warfare often targets. Source: MSB, “Supply readiness and critical infrastructure,” 2025; Wedin, KKrVA Proceedings and Journal, January 2025. |
| FACTBOX
Sweden's double exposure Sweden is one of the world's most trade-dependent countries, while its defense capabilities appear to reflect this dependence to a limited extent. According to available information, Sweden has 572 merchant ships at its disposal — none of which are container ships and only 121 fly the Swedish flag. Only flagged ships can be requisitioned in the event of war or the threat of war. Sweden contributed a staff officer to Operation Aspides — the EU's naval protection operation in the Red Sea, established in February 2024 partly to protect trade flows on which the Swedish economy is also dependent. In addition, a very large proportion of the active pharmaceutical ingredients used in Sweden are imported, according to several estimates around 80 percent, of which a significant proportion – often estimated at around 40 percent – comes from China or India. Against this background, a serious disruption to global supply chains does not appear to be a purely hypothetical scenario for Sweden, but rather as a specific risk profile that still seems to lack a clear owner in the planning of total defense. Source: Swedish Transport Agency, Ship Register 2025; EU EUNAVFOR Aspide mandate description, 2024; National Board of Health and Welfare, Pharmaceutical Supply and Preparedness, 2024. |
One of the fastest-eroding links was specifically brought up in the conversation — artificial intelligence:
| AI and AGI — the next driver of system collapse
Lars-Erik Lundin identifies artificial intelligence as the subsystem whose erosion is moving fastest and whose consequences are least understood. STARTING POINT At Davos 2026, Anthropic CEO Dario Amodei and Google DeepMind CEO Demis Hassabis publicly discussed the time horizon for AGI — Artificial General Intelligence. The estimates were 2028–2030. Europe is far behind. KEY POINTS • Autonomous weapons systems. Ukraine's defense minister — 35 years old, no military background, appointed for his AI expertise — is now operating autonomous weapons systems that decide when and who to attack. The international law thorny issue is unresolved. • Cyber threats. AI is accelerating cyber threats to critical infrastructure to a level where the traffic lights are already red. The threats are potentially existential for individual countries. • The balance of power. If China or the United States achieve superior AI capabilities, it could enable control of third-country systems from within — a new type of power that is not kinetic but equally crucial. CONCLUSION AI is not primarily a productivity tool — it is a strategic capability leap whose military, international law, and system-stabilizing consequences are not yet understood by decision-makers. From a systems perspective, this means that AI not only affects military capabilities, but also the information and decision-making systems on which international regimes rest. Source: Conversation between Lars-Erik Lundin and Michael Sahlin, 2026-05-29, 50:12–53:26 |
The conversation focused on what this means for a small state's room for manoeuvre:
| Proactivity and reactivity — the small state dilemma
Sahlin articulates the central tension for a small state facing global systemic risks: limited ability to influence, but inescapable obligation to protect. STARTING POINT Sweden and the Nordic countries can to a limited extent influence the global factors that drive systemic collapse. But that does not absolve us from the question of reactive protection — what can we do to protect ourselves when it happens? KEY POINTS • Proactive. Sweden can contribute to NB8, the EU and NATO with specific comparative advantages: the institutional presence of the total defense, underwater capabilities, LKAB deposits. But the leverage is limited the more global the issues are. • Reactive. The obligation to prepare reactive protective measures remains regardless of proactive capabilities. The hedgehog metaphor (neutral, waiting) is not available after EU and NATO membership. • The European level. EU and NATO membership means that the European level is the relevant field of action. The more the transatlantic divide falters, the more important the European level is as a replacement. CONCLUSION Sweden has chosen to be involved — in the EU and NATO — in order to have influence rather than just protect itself. The chosen strategy now requires that the EU level's ability to make decisions during a crisis be prioritized. Source: Conversation between Lars-Erik Lundin and Michael Sahlin, 2026-05-29, 1:14:02–1:14:56 |
What we cannot choose to do away with
Sweden and the Nordic countries are hardly spectators to a possible global system collapse process. We are very much part of it, exposed via trade chains, energy dependencies, data cables and the security policy framework that is a prerequisite for our freedom of action.
The analytical challenge is to try to take in the whole without drowning in it — to see what is happening in the Middle East, in Washington, and in Beijing not as distant abstractions, but as variables in our own planning horizon.
In our assessment, the most relevant planning assumption for Sweden in 2026 is likely not a third world war, but a gradually changed system state: where the institutions we have relied on provide increasingly poor service, where the role of the United States is no longer the same, and where our ability to make the right decisions at the right time may prove to be crucial.
This is, as we see it, a threat we need to take more seriously in our strategy.
In conclusion, the conversation underlined the need for a comprehensive, cross-border analytical capability:
| The danger of stupor — for a collective analytical ability
Lars-Erik Lundin ends the conversation with a warning against overly specialized analyses and an argument for the role of academia in the public debate. STARTING POINT Sweden, the EU and NATO produce a lot of analyses. But they are often sectoral: energy, cyber, migration, military. The global system collapse perspective requires horizontal integration — not more drainpipes. KEY POINTS • Planning assumptions change. 2022: the threat of invasion. 2023: NATO accession. 2025: Trump. Each new situation has replaced 'self-evident' assumptions. There must be an ability to systematically problematize them. • The Iran War's ramifications. The war against Iran affects Sweden via Ukraine, energy prices, refugee flows and the Middle East's connections to European security. Ignoring what is happening there is a planning error. • The added value of academia. The intelligence services have specific missions. The state administration has downpipes. The free academy — KKrVA — can provide the independent, interdisciplinary holistic perspective that is missing. CONCLUSION A collective public analysis capacity for global systemic risks and their consequences for Sweden is not a luxury phenomenon — it is part of civil defense. Source: Conversation between Lars-Erik Lundin and Michael Sahlin, 2026-05-29, 1:15:22–1:17:09 |
Sources
SIPRI: SIPRI Yearbook 2025 — Armaments, Disarmament and International Security. Stockholm International Peace Research Institute.
World Economic Forum: Global Risks Report 2026. Davos.
Farrell, H. & Newman, A.: “Weaponized Interdependence: How Global Economic Networks Shape State Coercion.” International Security 44 (1), 2019.
Jones, CI: “Intermediate Goods, Weak Links, and Superstars.” American Economic Journal: Macroeconomics 3 (2), 2011.
Jones, CI & Tonetti, C.: "AI and the Future of the US Economy." NBER Working Paper 34779, March 2026.
European Commission: ReArm Europe — Defense Investment Plan (2025).
European Commission: EU Action Plan for Cable Security (2024).
ENISA: Threat Landscape 2025. European Union Agency for Cybersecurity.
Directive (EU) 2022/2555 (NIS2 Directive) on measures for a high common level of cybersecurity.
IMF: “Shipping Cost Shocks and Global Supply Chains.” IMF Working Paper, 2024.
MSB: Supply readiness and critical infrastructure (2025). Swedish Civil Contingencies Agency.
National Board of Health and Welfare: Pharmaceutical supply and preparedness (2024).
Swedish Transport Agency: Ship register (2025).
EU EUNAVFOR: Operation Aspides — mandate description (2024).
Wedin, L.: “Flow security — a concept for our time.” Proceedings and Journal of the Royal Academy of Military Sciences, January 2025.
The Defence Committee: Gathering forces — The direction of Sweden's defense and security (Ds 2023:19).
EUI (European University Institute): European Crisis Management Capacity Review 2026. Florence.