Insurance that increases your exposure is bad insurance. Yet that is how the global security economy now behaves. The headline cost of violence reached about 19.1 trillion dollars last year, as deaths climbed to a 25-year high and militaries rearmed. That number looks like protection. It reads like fragility. Military outlays rose by roughly 540 billion to 9 trillion. Internal security spending hit about 5.7 trillion. Conflict-linked GDP losses surged 44 percent to around 462 billion. The bill for refugees and displaced people is roughly 343 billion, with 122 million uprooted. These are not just costs. They are signals that the system is operating closer to its stress limits, with fewer buffers and more ways to fail.
Markets still worship compounding. Governments now budget against it. Redirect enough capital from high-return civilian uses and the long arc of productivity bends down. The rearmament cycle is broad: Europe is pouring billions into defense; Japan plans to lift defense to two percent of GDP; the United States already spends close to 949 billion; China is near 450 billion in international dollars. Eighty-four countries raised military outlays as a share of GDP. This is not transient. It is a repricing of priorities that bleeds into interest costs, taxes, and the opportunity set for private investment.
Think of growth as a flywheel. It accelerates when education, infrastructure, and research get consistent fuel. Defense spending can seed useful technologies, but the yield is lumpy and path-dependent. The 540 billion step-up in one year is material compared with global R and D. When it is debt-financed in aging societies with low trend growth, the marginal dollar likely carries a higher risk premium. Over time, a budget that prioritizes armor over network effects crowds out the very compounding that pays for security tomorrow.
In game theory, the arms race is a prisoner’s dilemma: rational players escalate even though mutual restraint would make them better off. Accounting won’t show the foregone outcome. Deterrence is inherently invisible when it works, so the feedback is noisy. We then measure what we can: readiness scores, sortie counts, inventories. Goodhart’s Law kicks in. When a measure becomes a target, it stops being a good measure. The incentive becomes more hardware, not necessarily more security.
Look at the second-highest line item: internal security, roughly 5.7 trillion. That is policing, courts, and the architecture of domestic order. It is a sign that we insure against each other as much as against an external foe. The cost can buy calm, or it can buy the appearance of it. If the goal is public trust and social stability, you are paying for a stock that comes from flows of opportunity. The ledger records the expense, not whether it reduces the variance of future outcomes. A fragile equilibrium is expensive to maintain and quick to unravel.
Conflict losses do not scale linearly. They arrive in lumpy, fat-tailed distributions. The reported 44 percent jump in GDP losses last year, and the fourfold increase since 2008, fits that pattern. You can model expected losses and assign a discount rate. But you cannot wish away variance. Robust systems carry slack and redundancy to absorb shocks. Budgets that run hot, with rising mandatory outlays and rate sensitivity, lose that slack.
This is engineering 101. Build a bridge at ninety percent of its load limit, and small surprises cause collapse. The same applies to fiscal capacity. Layer defense increases on top of rising health and pension costs, then add debt service in a higher-rate world, and the buffer disappears. When the improbable happens, policymakers reach for the only lever they can pull quickly: taxes, austerity, or inflation. Markets respond by demanding more compensation for risk. The insurance premium rises because the deductible just did.
Defense looks like steel and missiles. In practice it is semiconductors, materials, machining, energy, logistics, and software. The supply chain is concentrated and slow to scale. Sole-source components, treaty constraints, and export controls add friction. Rebuilding munitions output in Europe will take years. Lead times stretch. The era of just-in-time inventory left thin buffers for a just-in-case world.
This is the second paradox. We are spending more to buy time we cannot. The arsenal narrative assumes production elasticity that no longer exists. Rare earths, specialty chemicals, high-end chips, and propulsion systems sit behind chokepoints. Shipping lanes and grids are vulnerable. Even the trained labor to run the lines is scarce. That is not a procurement problem; it is a system design problem. War plans that assume instant surge capacity are promises the industrial base cannot keep.
Defense equities have rallied, aided by retail flows and a search for earnings stability in an unstable world. The narrative is simple: budgets will rise, orders will follow, cash flows will expand. Some of that is true. But markets are extrapolating straight lines from curved realities. Governments squeeze suppliers when margins look rich. Cost-plus contracts cap profitability. Export approvals get politicized. Windfall optics trigger regulation and taxes. ESG screens may be loosening, but reputational risk has not vanished. Even supply chain gains can be confiscated in renegotiations.
Investors also overrate the persistence of wartime multiples. The performance of defense stocks clusters around regime shifts and procurement cycles, not forever. If the growth is in low-margin munitions rather than complex platforms, the mix changes. If insolvency risk rises elsewhere in the economy due to higher rates and tighter fiscal space, cost of capital rises for everyone, including contractors. Pricing war like a secular growth sector mistakes a risk hedge for a forever moat.
The least discussed line item in the conflict bill is the one that compounds on its own: social fracture. The world has 122 million displaced people. Refugee and IDP costs were estimated at around 343 billion last year. That is not just a humanitarian crisis; it is a supply chain, housing, and labor market shock that crosses borders and lasts years. Internal security spending then grows to manage symptoms, not causes.
A contrarian view says the most effective deterrent is a society that works. A budget that cuts prevention to fund response increases long-run security costs. If defense displaces education, affordable housing, and infrastructure, you widen the funnel of grievance. Then you pay for more policing and prisons. History is blunt on this trade. Late-imperial Rome financed legions while hollowing the middle. The United States tried to fund both the Great Society and a war in Southeast Asia; inflation followed and trust eroded. You do not need to agree on policy to see the mechanism. Fragility sells protection. Protection, if misallocated, creates more fragility.
The inversion is simple: stop buying only armor, start buying shock absorbers. In practical terms, that means redundancy, modularity, and slack across energy, food, health, and digital infrastructure. Harden grids. Diversify supply chains for critical inputs. Hold strategic inventories. Invest in dual-use technologies where civilian spillovers are large. Train for civil defense and recovery, not just expeditionary operations. Create de-escalation channels that reduce error risk because most wars start with miscalculation.
In probability terms, reduce variance at the source rather than paying to contain it at the boundary. In game theory terms, increase the payoff to cooperation relative to defection by lowering the cost of verification, communication, and credible restraint. Deterrence still matters. But a portfolio that only buys out-of-the-money puts will bleed premium forever and still blow up on a gap. The goal is a system that gains from stress, not one that looks strong right up to the moment it fails.
The 19 trillion headline is not the story. The story is how much of that spending lowers resilience elsewhere, how much of it buys signaling instead of capability, and how little of it is aimed at the knots that actually tighten in a crisis. Are we building gold-plated armor while letting the foundations crack. The bill suggests we are. The risk is that the next shock is not a test we pass, but a stressor that reveals we mistook cost for safety.
