This page holds the text of every public page on worldwar3.army in one place so crawlers, search engines and AI agents can read and cite it. Machine-readable table: agent-data.json and agent-data.csv. Summary for language models: llms.txt. Updated 2026-10-26.

Page table

Page text

Relative Decline Index: 20 Nations Scored | worldwar3.army

Source: https://worldwar3.army/

Which great powers are declining, and how fast? 20 nations scored on power level and decline velocity, plus the conditions that turn decline into war.

Nations do not fail. They get out-grown. The United States produced more in 2026 than in any year of its history, and its share of the world fell anyway. That is the whole of relative decline in one sentence — and it means the only question worth asking is not whether a power is declining, but how fast, in which direction, and whether the speed is changing. This site measures that. Twenty nations, five engines, two axes: where a power stands, and how quickly it is moving. Then it prices what happens when several large powers reach the same conclusion about their own windows at the same time. US share of world output 14.5% 2026, current prices. The 1986 peak was 22.3%. Nothing got smaller — the rest of the world got bigger. China population change −3.39M 2025 — the fourth consecutive annual fall, to 1.405 billion. Russian births, Q1 −6% 2026. Roughly 272,000 — a two-hundred-year low, before counting the war dead. State-based conflicts 65 2025. A record since records began in 1946. Interstate conflicts doubled for the second year running, to eight. In short Who is declining fastest, and does it mean war? On the Relative Decline Index (v1.0, scored 22 September 2026), Russia is declining fastest, at a velocity of −5.6, and the United States slowly, at −1.4. Decline raises the risk of conflict only when specific conditions line up, which is why the site prices those conditions and does not predict a world war. Updated 19 Oct 2026 Key facts As of the 22 September 2026 scoring, the index ranks 20 nations on power level (0–100) and decline velocity (−10 to +10); the next scheduled review is 31 December 2026. As of that scoring, the fastest decline velocities are Russia (−5.6), Italy (−4.0), Germany (−3.8), China (−3.4) and South Korea (−3.2). As of that scoring, the fastest risers are Vietnam (+4.2), Indonesia (+3.8), India (+3.6) and Poland (+3.1). China is scored as peaking rather than collapsing, with a power level of 74 and a velocity of −3.4; the United States has a power level of 88 and a velocity of −1.4. The Uppsala Conflict Data Program recorded 65 state-based conflicts in 2025, the most since its records began in 1946, including eight interstate conflicts (up from two in 2023). China’s population fell by 3.39 million in 2025, to about 1.405 billion, the fourth annual fall in a row (National Bureau of Statistics figures as reported by Caixin, 19 January 2026). Go deeper: the five conditions and the flashpoint board , the five engines of decline , who outlasts whom , and the chapters on China , Russia , Europe and the United States . The instrument Level is where you stand. Velocity is how fast you are losing it. Every discussion of decline collapses these two into one word and destroys the information. A power shedding a tenth of a point of world share each year is in a two-century glide. A power shedding two points a year is in a rout. Both get called “decline.” So the index separates them. The horizontal axis is power level — a composite of output share, manufacturing share, deployable force, technology-frontier presence, financial centrality and alliance depth. The vertical axis is decline velocity — half the annualised five-year change in that level, half a forward score built from the demographic and fiscal commitments already locked in. Each marker carries a tail showing where it sat five years ago. Power level against decline velocity, 2026 Click any marker for the full chapter. Hover for the score. Top right Incumbent, holding High power, flat or positive velocity. The United States sits here thinly, on the strength of migration, the technology frontier and the dollar. India is climbing toward it. Bottom right Incumbent, slipping High power, negative velocity. China and Germany. This is the most dangerous quadrant in the model, because it contains powers with the capability to act and a reason to act soon. Top left Rising Modest power, positive velocity. Indonesia, Vietnam, India, Poland, the Gulf. Australia holds its position at small scale, which at this point in the cycle counts as winning. Bottom left Spent Modest power, negative velocity. Russia, Italy, South Africa — powers whose influence is outliving the material base that produced it. The shape Decline is not a slope. It is a sequence. Powers do not descend at a constant rate. They plateau, panic, rebound and step down. America declined hard through the 1970s and then had its best fifteen years. Russia declined catastrophically, genuinely recovered on hydrocarbons for a decade, and is now inside a second and worse decline. That second case deserves its own stage. A second life is a completed decline followed by a real rebound — and it is the most dangerous position in the model, because the first decline consumed the asset that a second recovery would require. In Russia’s case that asset was people, and they are not coming back. The chapters Eight arguments, each with its own counter-case. Every chapter runs the same spine: the number that frames it, the mechanism, the strongest case against, what would falsify it, and what to watch with dates attached. The counter-case is not a formality. It is the difference between research and doom-blogging. The thesis the domain promises Decline becomes war through window logic, not desperation. A power whose position is deteriorating on a known schedule faces a choice between acting while it still can and accepting a permanently worse settlement later. Demographic decline is uniquely dangerous here because it is the most legible kind: a leadership can read the birth cohorts and know, to the year, when the option closes. Five conditions convert decline into fighting, and each is independently observable. The board below scores every flashpoint against all five. Probability bands are subjective five-year estimates, published with resolution dates and graded afterwards in public. What this site is not claiming Not that a world war is coming. The claim is narrower and testable: the conditions that have historically preceded great-power war are present in more places, simultaneously, than at any point since 1945. That is a statement about hazard rates. In 2025 the world recorded 65 state-based conflicts, eight of them interstate, and 244,600 deaths from organised violence against roughly 187,000 the year before. The trend is the evidence; the outcome is not written. Read the full conflict analysis, the five conditions, and the forecast ledger → Latest Velocity Report What moved this week Russia's War Budget Meets a 14% Interest Rate, October 2026 — Russia plans 17.1tn roubles of defence spending in 2027, 27% above plan, as the central bank holds at 14% and lifts its inflation forecast to 7.1-7.6%. 26 Oct 2026 Grey-Zone Velocity: South China Sea, October 2026 — Philippine monitoring detected 101 Chinese vessels in September 2026, nearly double August's count. Six confrontations in ten days. Velocity, not presence. 12 Oct 2026 Standards An index whose workings [continues on the page]

The five engines of decline | The Relative Decline Project

Source: https://worldwar3.army/theory.html

Kennedy, Gilpin, Organski, Tainter, Turchin and the demographers — five traditions, five engines, and the insight that the fast engines fail first while the slow engines decide the outcome.

The framework Five engines. Five half-lives. Decline is not one process. It is five, running at different speeds, and a nation’s trajectory is their vector sum. Getting this right is the difference between “everyone is declining” — which measures nothing — and a model that tells you which declines are recoverable. Engine What it measures Half-life Recoverable? Demographic Births, working-age trajectory, dependency ratio, net migration 40–80 years Only through migration, and only with consent Fiscal Debt path, interest burden, reserve privilege, external balance 10–25 years Yes, politically expensive Productive Productivity growth, investment efficiency, industrial base, energy cost 10–20 years Yes, with capital and time Institutional State capacity, rule of law, succession clarity, policy reversibility 5–30 years Slowly, and rarely under stress Cohesive Trust, polarisation, inequality, elite overproduction, legitimacy 2–15 years Yes — this one moves fastest in both directions The fast engines fail first. The slow engines decide the outcome. That is the whole insight, and almost every confused argument about national decline comes from ignoring it. Cohesion cracks in a decade and produces the headlines — polarisation, protest, institutional distrust, a surprising election. It is also the most recoverable. Demography fails over half a century, produces almost no headlines at all, and is effectively irreversible on any policy timescale. So a country whose demographic engine has already seized — China, Russia, Italy, Korea, Japan — is on a fundamentally different road from one whose engines are merely misfiring, however similar the news coverage looks this year. The United States has a cohesion score of 34, which is genuinely bad and genuinely alarming. It also has a demographic score of 62, which is why it sits in the holding quadrant and Italy does not. The five traditions Paul Kennedy — imperial overstretch (1987) Great powers decline when the military commitments accumulated at their peak outlive the economic base that paid for them. The test is the ratio of security spending to productive investment. The critique, which is devastating and should be quoted openly: Kennedy called American decline at the precise moment America was about to have its best fifteen years. The Soviet Union collapsed, the productivity boom arrived, and the debt ratio fell. It is the single best argument for humility in this entire genre, and any site making Kennedy’s argument owes the reader that disclosure on the first page. Why the idea survives the bad timing: the mechanism has arrived by a route Kennedy did not anticipate. The United States now spends more servicing its debt than defending itself — not because the garrisons got expensive, but because the borrowing that paid for them did. That is overstretch with a financial transmission rather than a military one. Robert Gilpin — the cost of order (1981) Hegemony is a business. The hegemon supplies public goods — sea lanes, a reserve currency, a settlement system, extended deterrence — and collects rents: seigniorage, standard-setting, alliance leverage. Decline begins when the cost of supplying the goods rises faster than the rents. The observable is beautifully simple: is the leading power adding or shedding commitments? France’s African base network collapsing is a Gilpin ledger entry. Russia’s Africa Corps evacuating Kidal is another. The American debate over war powers is a third. Every one of them is a power discovering that a commitment costs more than it returns and quietly closing the position. Organski and Kugler — power transition (1958/1980) War is most likely not when a challenger is weak, nor when it is dominant, but during crossover — and only when the challenger is dissatisfied with the existing order. Graham Allison popularised this as the Thucydides Trap in 2017. The critiques are strong and this site carries them in full on the conflict page : determinism, non-uniform measurement, the difficulty of operationalising “satisfaction,” the peaceful counterexamples, and the self-fulfilling-prophecy risk. The position here is that power transition identifies a hazard rate rather than a destiny, and that the dissatisfaction variable — not the power variable — is the one worth watching. Joseph Tainter and Peter Turchin — complexity and elite overproduction Tainter (1988): societies solve problems by adding complexity; complexity has declining marginal returns; eventually the society cannot afford the apparatus that keeps it functioning, and it simplifies — sometimes abruptly. Turchin adds the quantitative machinery and two measurable drivers: popular immiseration and elite overproduction — too many credentialed aspirants competing for too few elite positions, reliably producing counter-elites and factional conflict. The clearest live case in this index is China: youth unemployment at 16.9% among non-students in March 2026, produced by a graduate pipeline built for a decade of 8% growth meeting an employment mix that no longer generates those jobs. It is elite overproduction with Chinese characteristics, it has no electoral safety valve, and it is the most under-priced risk in the Chinese story. The demographic school Populations are the slowest-moving and most reliable of all the variables. A birth cohort that does not exist in 2026 cannot be drafted, taxed or employed in 2046. This is the index’s anchor, because it is the domain in which forecasting is least dishonest — the twenty-year-olds of 2046 are already a fact. One caution stated explicitly. Glubb’s “fate of empires” — the ten-generation cycle ending in an “age of decadence” — circulates widely in this subject area and is not scholarship. It is a retired general’s pattern-matching, unfalsifiable and heavily moralised. This site uses the demographic evidence, which is excellent, and does not adopt the moral frame, which is not evidence at all. The lifecycle Decline is not a slope but a sequence, and the stages are not interchangeable. The index places every nation on a six-stage track. Two stages deserve explanation because they are the model’s original contributions. Apex is scored as more dangerous than any decline stage. A power at or just past its peak has maximum capability and a deteriorating forecast — the exact combination that produces window logic. China and South Korea sit here for very different reasons. Second Life describes a power that completed a full decline, executed a genuine rebound, and is now inside a second and structurally worse descent. Russia is the only case in the index, and the classification carries a specific claim: the first decline consumed the asset a second recovery would have required. In Russia’s case that asset was people. This is why Russia scores the steepest velocity in the model at −5.6 despite a power level of only 31, and why the index treats it as the most volatile rather than the most powerful actor. What this framework refuses to do It does not produce a single decline number. A composite that averages Nigeria’s demographic 82 with its institutional 26 tells you nothing about either. The engines are published separately for exactly this reason. It does not treat decline as moral failure. A low engine score is a measurement, not a judgement. Europe trading global weight for stability and longevity may well be a [continues on the page]

Method — every weight, every source | The Relative Decline Project

Source: https://worldwar3.army/method.html

The full Relative Decline Index: axis definitions, engine weights, indicator library, confidence grading and the complete score table for all twenty nations.

Every weight. Every source. An index whose workings are hidden is a horoscope. An index whose workings are published is an argument — and arguments can be won or lost on the evidence. Everything needed to disagree with this one is on this page. The two axes Power level (0–100) Where the nation stands today. A composite of six components, weighted as shown. The United States anchors the upper range at 88 rather than 100, leaving headroom — an index whose leader is already at maximum cannot record a rise. Component Weight Principal indicators Share of world output 25% GDP share, PPP and nominal blended 50/50 (IMF WEO) Manufacturing value-added share 20% UNIDO / World Bank MVA share of world total Military capability 20% Expenditure (SIPRI), deployable force, shipbuilding and munitions throughput Technology frontier 15% Semiconductor, AI compute, aerospace, pharmaceutical frontier presence Financial centrality 12% Reserve-currency share, settlement share, market depth Alliance depth 8% Treaty commitments weighted by partner capability Decline velocity (−10 to +10) Not a GDP forecast. The first derivative of relative position, constructed in two equal halves: 50% trailing. The annualised five-year change in power level, scaled so that a one-point annual loss of level reads as roughly −1.5. 50% forward. A momentum score built only from commitments already locked in — birth cohorts already born, pension liabilities already accrued, debt already issued, retirement ages already legislated. Nothing discretionary, nothing forecast. The forward half is what separates this index from a growth table. China’s trailing level rose between 2021 and 2026, from 70 to 74. Its forward half is deeply negative because the 2046 workforce is already born and already too small. Net: −3.4. A pure trailing measure would have scored China positive and missed the entire argument. The five engines Each engine is scored 0–100, where 100 is healthy . A low score is a failing engine, not a moral judgement. Engines are published separately and never averaged into a single figure, because averaging Nigeria’s demographic 82 with its institutional 26 destroys both numbers. Engine Indicators Primary sources Demographic TFR, working-age trajectory to 2050, old-age dependency, net migration rate, migration policy durability UN WPP, OECD, national statistical offices (NBS, Rosstat, Destatis, e-Stat) Fiscal Debt/GDP path, net interest as share of outlays, reserve-currency privilege, external balance, revenue concentration CBO, IMF Article IV, national treasuries, PGPF Productive Labour productivity growth, investment efficiency, industrial output vs peak, industrial energy cost, robot density OECD, ifo, IMF, national industrial series Institutional State capacity, rule of law, succession clarity, policy reversibility, regulatory predictability V-Dem, Freedom House, World Bank WGI Cohesive Institutional trust, polarisation, inequality, elite overproduction proxies (graduate unemployment), legitimacy indicators V-Dem, national labour statistics, survey series The full index All twenty nations, sorted by power level. Engine columns abbreviated: Demo, Fisc, Prod, Inst, Coh. Confidence grading Every figure on this site carries one of three grades, inherited from the underlying research methodology and applied consistently: High Official statistics, primary disclosures, treaty text, or two independent outlets in agreement. Medium Single credible source, reported-but-unconfirmed, or an official target rather than an outcome. Low Model output, long-horizon extrapolation, advocacy-cited figures, or this index’s own subjective judgement. Two disciplines follow from this and are applied throughout. First, claims sourced to a single outlet are labelled as such rather than presented alongside confirmed ones. Second, ranges are published as ranges. Russian war deaths appear as three separate figures on three separate bases — a named-and-confirmed floor of 223,539, a probate model at roughly 352,000, and a GCHQ assessment near 500,000 — because collapsing them into one number would misrepresent all three. Known limitations Stated plainly, because an index that does not publish its weaknesses is asking to be trusted rather than checked. The forward half of velocity is a construction, not a measurement. It is defensible and the inputs are locked-in facts, but the weighting between locked-in demography and locked-in fiscal commitments is a judgement call. Power level weights are argued, not derived. There is no objective basis for 25% output versus 20% manufacturing. Different weights move the horizontal positions; they move the vertical ones far less, which is the more important axis. Institutional and cohesion scores rest on indices with their own methodological disputes. V-Dem’s recorded 24% one-year fall in the US Liberal Democracy Index — and a rank move from 20th to 51st of 179 — is a large swing that reasonable analysts contest. Probability bands on the conflict board are subjective. They are the author’s judgement, graded Low confidence, and published only because unscored intuitions are worse than scored ones. Nations are unequal units. The EU is disaggregated into Germany, France, Italy, the UK and Poland because the dispersion within Europe is larger than the dispersion between several continents. That is an analytical choice with consequences. An unrevised index is a wrong index. Quarterly re-scoring, published changelog, with the previous version left visible. Revision policy The index is re-scored quarterly. Every revision publishes what moved, by how much, and on what evidence. Prior versions remain accessible. When a score changes because the world changed, that is an update; when it changes because the index was wrong, that is a correction, and it is labelled as one. Continue The five engines and the literature · Editorial standards and corrections · Back to the index

World War 3 and Relative Decline: The Theory

Source: https://worldwar3.army/conflict.html

Decline leads to conflict through window logic, not desperation. Five conditions, eight flashpoints, probability bands with resolution dates.

The thesis the domain promises World War 3 and a theory of relative decline. This is the most disciplined page on the site, and it has to be. A domain like this one earns trust only by refusing to do the thing the domain implies. So: no prophecy, no inevitability, no world war. Conditionals with named triggers, probability bands with resolution dates, and public scoring afterwards. EIGHT FLASHPOINTS · FIVE CONDITIONS EACH State-based conflicts, 2025 65 A record since UCDP records began in 1946. Interstate conflicts 8 Doubled for the second consecutive year, from two in 2023. Highest since 1946. Fatalities 244,600 2025, against roughly 187,000 in 2024. One-sided violence ~76,500 Up more than 400%. Highest since 1994. Updated 22 October 2026 The short answer This page does not predict a world war. It argues that relative decline raises the risk of conflict through “window logic”: a power whose position is deteriorating on a known schedule may act while it still can. Five observable conditions are scored against eight flashpoints, and every probability band is published with a resolution date. Key facts The Uppsala Conflict Data Program (UCDP) recorded 65 state-based conflicts in 2025, the most since its records began in 1946. Eight of those 2025 conflicts were interstate, up from two in 2023 and the highest count since 1946. UCDP counted about 244,600 deaths from organised violence in 2025, the second-deadliest year since the 1994 Rwanda genocide. UCDP recorded about 76,500 deaths from one-sided violence in 2025, an increase of more than 400% and the highest since 1994. UCDP counted 13 conflicts with at least 1,000 battle-related deaths in 2025, the most since 1992. The probability bands on this site are subjective five-year estimates of a state-on-state clash above 1,000 battle deaths, labelled Low confidence, and each is scored at its resolution date. The mechanism: window logic The popular model of decline-driven conflict is desperation — a failing state lashing out. That model is mostly wrong, and it misdirects attention toward the weakest powers when the risk sits with the strong ones. The better model is window logic . A power whose relative position is deteriorating on a known schedule faces a choice between acting while it still can and accepting a permanently worse settlement later. It is not desperate. It is doing arithmetic. And the arithmetic gets worse every year it waits, which converts patience from a virtue into a cost. Demographic decline is uniquely dangerous inside this model because it is the most legible form of decline there is. Fiscal trouble can be inflated away, argued about, or deferred. Productivity is contested and revised. But a leadership can read its own birth cohorts and know, to the year, when the manpower, the tax base and the strategic option all close. There is no ambiguity and no policy that reverses it inside a decade. This is why the bottom-right quadrant — high capability, negative velocity — is the most consequential cell in the index. It is not where the weak sit. It is where the strong sit with a clock. The five conditions Each is independently observable, and each flashpoint on the board below is scored against all five. The claim is not that five-of-five means war. It is that the historical base rate rises sharply with the count, and that a case meeting all five deserves a completely different level of attention from one meeting two. A closing window, visible to the leadership that holds it. Not merely decline — decline with a date on it. China’s military-demographic window across roughly 2027–2035; Russia’s manpower window, already half shut. A dissatisfied riser. This is the Organski variable and it is the discriminator that most commentary drops. The United States overtook Britain without war because it was a satisfied riser that accepted the existing order. Power transition alone predicts nothing; dissatisfaction is what makes transition dangerous. A contested fixed resource. Minerals, water, sea lanes, seabed, orbital slots. Fixed supply is what makes bargaining fail, because there is no version of the deal where both sides get more. An eroded guarantee. The defender’s commitment is doubted by both the challenger and the protected party. This is the classic deterrence-failure condition, and it is doubted symmetrically — which is why reassurance and deterrence pull against each other. Domestic legitimacy pressure. Diversionary incentives, which rise precisely as the cohesion engine fails. In the index, the cohesion engine is the fastest-moving of the five — it can crack in a decade — which makes it the one most likely to change a flashpoint’s score between quarterly reviews. The flashpoint board How to read the bands The probability bands are subjective five-year estimates of a state-on-state armed clash exceeding 1,000 battle deaths. They are the author’s judgement, not a model output, and they are labelled Low confidence throughout for that reason. Each will be scored at its resolution date and the record published — including the misses, prominently. Three points on interpretation that matter more than the numbers: High probability does not mean high consequence. The South China Sea carries a higher band than Taiwan and a far lower escalation ceiling. Hormuz carries the highest band on the board and is the case the world has already absorbed repeatedly. Read the band and the character together or you will read both wrong. The Sahel entry is not a forecast. It is already the largest fatality source in the dataset. It sits on the board because a site about great-power decline that only counts great-power casualties is measuring the wrong thing. Greenland is on the board at under 5% because acquisition contests run through licences, export credit and basing rights long before they run through force. Low probability of shooting is not low significance. The critiques, carried openly Power-transition theory and its popular form, Graham Allison’s Thucydides Trap, are the intellectual engine of this page, and the objections to them are strong enough that presenting the framework without them would be dishonest. Determinism. The framework is routinely read as predicting an inevitable clash. It does not, and this site does not. It identifies conditions under which war becomes significantly more probable. A hazard rate is not a destiny. Measurement. Allison’s sixteen-case file has been criticised for non-uniform measurement and selection effects, and the veneer of scientific rigour it carries is not matched by replicable standards. Worse, “satisfaction” — the variable that does most of the work — is genuinely hard to operationalise. Power is measurable in GDP and budgets. Contentment with the international order is not. Counterexamples. Several transitions have passed without war, most notably the United States overtaking Britain. Any honest use of the framework has to explain the peaceful cases, not just the violent ones. Self-fulfilment. The most serious objection. If policymakers believe war is fated, they take precautionary actions that are indistinguishable from preparation, and the other side responds in kind. A site that prices conflict has to be conscious that pricing is [continues on the page]

Who outlasts whom | The Relative Decline Project

Source: https://worldwar3.army/outlast.html

The historical record of who has already outlasted whom, and the forward ordering. Endurance is not velocity: a nation can fall fast and still outlast one falling slowly.

The question underneath all the others Who outlasts whom. Every other page on this site measures speed . This one measures endurance — and the two orders are not the same. A nation can be falling fast and still outlast one that is falling slowly, because falling and ending are governed by different engines. Before projecting forward, it is worth being exact about what the record actually shows. Most of what people call the death of a great power was not a death at all. First: what has already outlasted what The single most important correction to this whole genre is that in the modern record, nations almost never die. Regimes die, empires dissolve, and positions transfer. Of the great powers of 1900, essentially all still exist as states. What ended were the arrangements sitting on top of them. So “who outlasts whom” has to be asked at the right level, and there are four, with wildly different survival rates: Level What it means How often it ends Example Position Primacy, reserve currency, the right to set the rules Routinely — roughly once a century Britain to the United States, 1918–1945 Regime A ruling system and its governing bargain Often — several times a century The USSR, 1991 Empire Rule over territory that is not the core nation Often, and usually fast once it starts Austria-Hungary and the Ottomans, 1918 Nation / people The population and its continuity Very rarely, and almost only through demography Essentially none in the modern record The record, stated plainly Byzantium outlasted Rome by almost a thousand years. The Western empire fell in 476; the Eastern held until 1453. The half that survived was not the richer or more prestigious half — it was the one with defensible geography, a working tax system and an intact administrative class. This is the oldest and best-calibrated lesson available: fiscal and institutional competence outlasts wealth and prestige. The United States outlasted the Soviet Union — without fighting it. 1991. The transition that most defined the last century was decided by economic and demographic exhaustion on one side, not by battle. The United States also outlasted British primacy — and Britain is still here. The position transferred between 1918 and 1945. Britain lost the position and kept the country, which is the normal outcome and the one this site keeps insisting on. Russia outlasted the USSR. The regime died; the state continued. Important, because it is exactly the distinction that makes today's Russian question answerable: the question is not whether Russia ends, it is whether this Russia does. Japan outlasted its own empire, and got richer doing it. The empire died in 1945; Japanese living standards then rose for forty-five years. Losing a position is not the same as losing a future. Spain, Portugal and the Netherlands each lost global primacy and are now among the most prosperous places on earth. Every one of them would have been written off in the year they lost it. In every single case, the thing that ended was a position or a regime. The thing that persisted was a people. Any forward ranking has to respect that, or it is writing fiction. So what actually governs endurance Decline velocity is driven by the fast engines — cohesion cracks in a decade, fiscal room closes in twenty years, industry hollows in fifteen. That is what produces the headlines and the falling share. Endurance is governed by the slow engines, and above all by one of them. A regime can be replaced. A currency can be devalued. An industrial base can be rebuilt in a generation — Germany did it twice in one century. A generation that was never born cannot be recovered on any policy timescale at all. The endurance score therefore weights demography at 45%, institutional capacity at 25%, fiscal room at 20% and productive base at 10%, adds a population-mass bonus on a log scale, and applies a penalty where the recovery asset has already been spent — that is, to nations in second life or terminal decline. Full workings are on the method page . The endurance ladder Ranked by survival horizon, not by current power and not by how fast they are falling. Click any nation for its chapter. The five results that matter These are the pairwise calls the ladder actually makes. Each one is counter-intuitive against the way these countries are usually discussed, and each follows directly from the weighting above. India outlasts China Endurance 67.6 against 44.6 — the widest gap between two peers in the model. India’s demographic engine reads 74 to China’s 22. China is far more powerful today and will be for decades; that is a statement about level, not about horizon. On current cohorts this is not close. Nigeria outlasts Italy, Japan and Korea Endurance 59.2 against 26.8, 41.9 and 43.2. Nigeria has the weakest institutions in the model (26) and the strongest demographic engine anywhere (82). Every one of the three it outlasts is richer, better governed and more orderly — and running out of people. The ladder is brutal about which of those is recoverable. The United States outlasts every other incumbent Endurance 62.5, fourth overall and first among the major powers, on a demographic engine of 62 that exists entirely because of migration. Remove that one policy and America drops into the Japan band. The whole American endurance advantage rests on a single contested lever. Australia outlasts everyone Endurance 73.4, first in the model, on a population of 27 million. Small is not the same as fragile. Australia scores 68 demographic, 78 institutional and 66 fiscal — the only nation with no failing slow engine at all. It is the proof that endurance is a quality question before it is a size question. Almost everyone outlasts Russia Endurance 21.3, last in the model. Births at a two-hundred-year low, the war consuming the exact cohort that would have funded a recovery, an institutional score of 22 and no succession mechanism. The honest form of this claim: it is this Russian arrangement that is short-lived. Russia outlasted the USSR and will outlast this too. Poland rises but ranks 14th on endurance The cleanest demonstration that the two orders diverge. Poland has the third-fastest velocity in Europe at +3.1 and a demographic engine of 31. It is winning the decade and losing the century, which is a completely coherent thing for a country to be doing. Where the two orders disagree This table is the whole argument of the page in one place. The gap column is how many places a nation moves between the velocity ranking and the endurance ranking. A large positive gap means a nation is falling faster than it is ending . A large negative gap means it is ending faster than it is falling — the more dangerous condition, and the one almost nobody prices. The steelman Demography is not destiny, and a 45% weighting on it is [continues on the page]

United States decline: debt, interest and the dollar

Source: https://worldwar3.army/united-states.html

Is the US in decline? Net interest hit about $1.0tn in FY2026, above defence, while the dollar reserve share sits near 56%. Sourced, dated figures.

Chapter 01 · Incumbent, holding Decline by arithmetic, not by event. America is not collapsing. It is being out-grown — and it is spending the proceeds of its dominance on the interest on its dominance. Those are two different sentences about the same country, and both are true. Gross federal debt $39.39tn 6 July 2026. Growing roughly $7.7bn a day. Net interest, FY2026 $1.0tn Up from $375bn in 2019. Now past both defence and Medicare. Debt held by public 101% Of GDP. Reaches 120% by 2036, passing the 1946 record of 106%. Dollar share of reserves ~56% A 31-year low as central banks diversify into gold. Updated 23 October 2026 The short answer The United States is being out-grown, not breaking. Its share of world output is falling slowly while net interest on the federal debt reached about $1.0tn in FY2026, more than it spends on defence. The dollar’s reserve privilege, about 56% of global reserves, is what keeps $39tn of debt survivable. Key facts Gross federal debt was $39.39tn on 6 July 2026, growing roughly $7.7bn a day (Treasury / Joint Economic Committee). Net interest cost reached about $1.0tn in FY2026, up from $375bn in FY2019 (House Budget Committee / Peter G. Peterson Foundation). The CBO projects debt held by the public rising from 101% of GDP in 2026 to 120% in 2036, passing the 1946 record of 106%. Moody’s cut the US from Aaa to Aa1 on 16 May 2025, removing the last top-tier rating. China’s Treasury holdings were $693.3bn in February 2026, down about 42% from the November 2013 peak of $1.32tn; the UK ($897bn) is now the second-largest foreign holder (CRS / Statista). The dollar’s share of global reserves, about 56%, is at a 31-year low as central banks diversify into gold. The mechanism The American decline story is almost always told as a story about events — an election, a withdrawal, a bank failure. It is not. It is a story about arithmetic, and the arithmetic has two independent halves that happen to be arriving together. The first half is the denominator . The US share of world output has fallen from roughly 40% in 1960 to about a quarter by 2023, and on a current-price basis to around 14.5% in 2026 against a 1986 peak above 22%. Across 2017–2026 the average has been about a tenth of a percentage point a year. No American factory closed to produce that number. Asia simply got richer faster. The second half is the interest line , and this is the half that is genuinely new. In FY2026 the United States crossed a threshold no peacetime hegemon has crossed comfortably: it now spends more servicing its debt than it spends defending itself. Net interest hit roughly $1.0 trillion, a 2.5x increase in seven years, second only to Social Security among federal outlays. The Congressional Budget Office projects it exceeding Medicare by FY2028, all discretionary spending by FY2038, and becoming the single largest thing the federal government does by FY2048. This is Paul Kennedy’s imperial overstretch arriving by a route he did not anticipate. Not the cost of the garrisons — the cost of the borrowing that paid for them. The deficit is structural rather than cyclical: FY2026 runs $1.9tn (5.8% of GDP) on $7.4tn of outlays against $5.6tn of revenue, widening to $3.1tn (6.7%) by 2036. The One Big Beautiful Bill Act raised the ceiling by $5tn to $41.1tn in July 2025, scored by CBO at $3.4tn over ten years, by the Committee for a Responsible Federal Budget at $4.1tn, and by Cato at $6tn once dynamic interest effects are counted. Moody’s removed the last AAA in May 2025, warning of 134% of GDP by 2035. Why it has not broken Because of the privilege. The dollar still carries roughly 56% of global reserves and about 89% of foreign-exchange trades, and one credible estimate is that reserve-currency status lets the United States carry around 22% more debt than it otherwise could. That cushion is the entire reason $39 trillion is survivable. But the cushion is deflating, slowly. The dollar’s reserve share is at a 31-year low as central banks move into gold. China’s Treasury holdings are down roughly 42% from their November 2013 peak of $1.32tn to $693.3bn — a deliberate multi-year rundown, not a tantrum. The United Kingdom has quietly displaced China as the second-largest foreign holder at $897bn. And in March 2026, something instructive happened. Japan sold about $47bn of Treasuries and China about $41bn — not out of hostility, but to raise dollars to defend their currencies after the Iran war spiked oil. A resource conflict in the Gulf reached into the hegemon’s funding market inside a month. Total foreign holdings still stood near $9.5tn, up ~6% year on year, so this was not a stampede. It was a demonstration of a transmission channel. The evidence Figure Value Date Source Confidence Gross federal debt $39.39tn 6 Jul 2026 Treasury / JEC High Debt growth rate ~$7.71bn/day 2026 JEC / Treasury Fiscal Data High FY2026 deficit $1.9tn (5.8% GDP) FY2026 CBO Budget Outlook High Deficit, 2036 $3.1tn (6.7% GDP) proj. 2036 CBO Medium Debt held by public 101% → 120% GDP 2026→2036 CBO Medium Net interest, FY2026 $1.0tn FY2026 House Budget Cttee / PGPF High Net interest, FY2019 $375bn FY2019 PGPF High Avg rate, marketable debt 3.411% Jun 2026 Treasury High Moody’s downgrade Aaa → Aa1 16 May 2025 Moody’s / CNN High Japan Treasury holdings $1.24tn Feb 2026 CRS / Statista High UK holdings (now #2) $897bn Feb 2026 CRS / Statista High China holdings (now #3) $693.3bn Feb 2026 CRS / Statista High China holdings vs 2013 peak −42% 2013→2026 CRS Medium Mar 2026 official selling JP ~$47bn / CN ~$41bn Mar 2026 CNBC / Wolf Street Medium Dollar reserve share ~56% (31-yr low) 2026 PGPF / Wolf Street Medium US share of world GDP ~14.5% (current prices) 2026 IMF-derived series Medium The five engines America’s profile is unusual and worth reading carefully: a strong productive engine, an adequate demographic engine, and a fiscal engine in genuine distress — with cohesion the weakest reading in the model outside the authoritarian cases. The demographic score of 62 is the single most important number on this page, and it is why America sits in the holding quadrant rather than the slipping one. The United States is the only large rich economy still growing its population, and it does so through migration — which is precisely the lever Japan, Korea, Italy and China either cannot pull or will not. The cohesion score of 34 is the reason that lever is politically contested, and the two facts are not independent. The steelman The best case against this chapter is that it has been made before, by better analysts, and was wrong. Paul Kennedy published The Rise and Fall of the Great Powers in 1987 and called American overstretch at the precise moment America was about to have its best [continues on the page]

China: the decline nobody has priced | Decline Project

Source: https://worldwar3.army/china.html

China's decline began before its peak was recognised, and it is running across four systems at once: economy, demography, governance and social contract.

Chapter 02 · Apex The decline nobody has priced. China’s decline began before its peak was recognised, and it is running across four systems simultaneously — the economy, the demography, the governance structure and the social contract. This is the flagship contrarian call of the index, and it is built entirely on China’s own official data. The short answer · updated 20 October 2026 Is China in decline? It is peaking, not collapsing. On the Relative Decline Index China keeps a high power level (74) but a negative decline velocity (−3.4), because its demography, property sector and local finances are all deteriorating at once while its manufacturing engine is still strong. The index prices conditions, not outcomes: this is a statement about how fast China’s options are narrowing, not a forecast of what Beijing will do. Key facts, dated and sourced China’s power level is 74 and its decline velocity is −3.4 on the Relative Decline Index (v1.0, scored 22 September 2026); its engine scores are productive 71, institutional 44, fiscal 38, cohesion 37 and demographic 22 ( method and full score table ). China’s population fell by 3.39 million in 2025 to about 1.405 billion, the fourth straight annual fall (National Bureau of Statistics, via Reuters, 18 January 2026). China recorded 7.92 million births and 11.31 million deaths in 2025; births were 9.54 million in 2024 (National Bureau of Statistics, via Reuters, 18 January 2026). Real-estate investment fell 16.2% year on year in January–May 2026, with new starts down 22.6% and completions down 23.4% (National Bureau of Statistics). Youth unemployment among non-students aged 16–24 was 16.9% in March 2026 (National Bureau of Statistics). Local land-sale revenue fell from ¥8.49 trillion in 2021 to ¥4.15 trillion in 2025, and the IMF assesses roughly one in three local government financing vehicles as technically insolvent (see the evidence table below). Population change, 2025 −3.39M To 1.405bn. The fourth consecutive annual fall. Consecutive deflationary quarters 10 The longest stretch since the reform era began. Real-estate investment −16.2% Jan–May 2026, year on year. New starts −22.6%. Youth unemployment 16.9% March 2026, ages 16–24 excluding students. Four systems, one direction The mistake most analysis makes is treating these as separate stories with separate timelines. They are one story, and they compound. The demographic engine has seized China’s population fell by 3.39 million in 2025, the fourth straight annual decline. The working-age population peaked at around 990 million in 2020; the workforce has contracted roughly 6.8% since 2011, and young entrants are down 15% since 2015. The over-60 cohort stands at about 310 million (22%) and reaches roughly 402 million (28%) by 2040. China is making the 15%-to-25% aging transition in about two decades — Japan took three, and Japan was rich before it got old. The IMF puts aging alone at roughly two percentage points off annual GDP growth to 2050, with the growth path stepping down from 5.0% (2025) to 4.5% (2026) to about 3.5% by 2030 — some 2.7pp below the 2016–18 baseline. Pension spending goes from 5.4% to 15.3% of GDP by 2050 without reform (11.9% with it), and the main fund depletes around 2035. The policy response so far is a childcare subsidy of 3,600 yuan per child per year reaching roughly 20 million families. Against urban childcare costs, that is not a lever. It is a gesture. The property system is being deflated on purpose, and it is not finished Evergrande’s peak debt exceeded $300bn; against roughly $45bn of claims filed, about $255 million had been recovered by July 2025 — under a cent on the dollar. Country Garden restructured $17.7bn offshore. Vanke, the state-linked developer that was supposed to be safe, is being kept alive by Shenzhen Metro with more than ¥9.4bn of bonds maturing in six months. The flow numbers are worse than the headlines: real-estate investment down 16.2%, new starts down 22.6%, completions down 23.4% (Jan–May 2026), 100-city secondary prices still falling month on month, and roughly 20 million presold homes unfinished — families still paying mortgages on buildings that do not exist. The transmission into the state is the part that matters. Local land-sale revenue collapsed from ¥8.49tn in 2021 to ¥4.15tn in 2025. Hidden local debt exceeds ¥60tn (~$8.5tn), and the IMF assesses roughly one-third of local government financing vehicles as technically insolvent. The 2026 budget — ¥30tn of spending, an 8.1% broad deficit ratio, ¥6tn allocated to hidden local debt — is Beijing hauling the problem onto the central balance sheet. That is a solvency answer. It is not a growth answer. The governance structure is centralising into the problem Xi has swung the pendulum decisively toward centralisation: bond quotas set centrally, project selection set centrally, the anti-corruption apparatus institutionalised as a permanent disciplinary instrument. This is rational as crisis management and corrosive as an information system. A structure that punishes bad news does not receive accurate news, and a country running a controlled deflation of its largest asset class needs accurate news more than most. The social contract is quietly being renegotiated Youth unemployment at 16.9% among non-students is not a cyclical number. It is a structural mismatch between a graduate pipeline built for a decade of 8% growth and an employment mix that no longer produces those jobs. Peter Turchin has a name for this: elite overproduction — too many credentialed aspirants for too few elite positions, reliably producing counter-elites and factional conflict. It is the single most under-priced risk in the Chinese story, because it has no electoral safety valve and no obvious policy fix. The bargain was growth in exchange for control. Growth is now 4.5% and falling, and control is the only half of the deal the state can still deliver. The evidence Figure Value Date Source Confidence Population change −3.39M → 1.405bn 2025 NBS via SCMP High Working-age peak ~990M (2020) 2020→ CNN / CFR Medium Workforce contraction ~6.8% since 2011 2011–25 CFR Medium Over-60 population ~310M (22%) → ~402M (28%) 2024 / 2040 CFR Medium IMF growth path 5.0% → 4.5% → ~3.5% 2025–30 IMF Article IV Medium Aging drag on growth ~2pp per year 2024–50 IMF Working Paper Low–Med Pension spending 5.4% → 15.3% GDP to 2050 IMF WP Low–Med Pension fund depletion ~2035 proj. CFR / Merics Medium 2026 growth target 4.5–5% Mar 2026 NPC via CNBC High Deflationary pressure ~10 consecutive quarters to 2026 Asia Society Medium Youth unemployment (ex-students) 16.9% Mar 2026 NBS High Real-estate investment −16.2% YoY Jan–May 2026 NBS via Asia Times High New starts / completions −22.6% / −23.4% Jan–May 2026 NBS High Unfinished presold homes ~20 million 2026 est. IRR / WSJ Medium Local land-sale revenue ¥8.49tn → ¥4.15tn 2021–25 Observatorio Global Medium Hidden local debt >¥60tn (~$8.5tn) early 2026 IMF / Observatorio Medium LGFVs technically insolvent ~1 in 3 early 2026 IMF Medium Evergrande recovery rate ~$255m on ~$45bn claims to Jul 2025 SCMP Medium The five engines The shape of China’s profile is the argument: a productive [continues on the page]

Russia — the second life, ending | The Relative Decline Project

Source: https://worldwar3.army/russia.html

The only major power currently living through its second decline — and the first one is why the second is terminal.

Chapter 03 · Second life The second life, ending. Russia is the only major power currently living through its second decline — and the first one is the reason the second is terminal. The collapse of 1989–1999 consumed the demographic base that a second recovery would have required. What was spent cannot be spent again. The short answer · updated 12 October 2026 Is Russia in decline? Yes, and on this index it is the fastest-declining major power: a decline velocity of −5.6 on a power level of 31. The first collapse (1989–1999) spent the demographic base a second recovery would need, and the war is now drawing on the same age band. The index prices conditions, not outcomes: this is a statement about how fast Russia’s options are narrowing, not a forecast of what it will do. Key facts, dated and sourced Russia’s decline velocity is −5.6 and its power level is 31 on the Relative Decline Index (v1.0, September 2026); its demographic engine scores 14 out of 100 ( method and full score table ). Russia’s natural population decline in 2024 was 596,200, a 20.4% deterioration on 2023 (Rosstat via Moscow Times). Russia recorded about 272,000 births in Q1 2026, down 6% year on year, with a total fertility rate of 1.374 (Moscow Times; IPS Journal). Mediazona had named and confirmed 223,539 Russian war dead, including 7,345 officers, by 3 July 2026; this is a floor, not an estimate. Russia’s federal budget deficit was RUB 5.186 trillion (2.2% of GDP) in January–September 2026, and the Finance Ministry has raised its full-year forecast to RUB 7.345 trillion (3.2% of GDP) from a budget-law plan of 1.6% (preliminary data, Interfax, 9 October 2026). Russian oil and gas revenue in January–September 2026 was RUB 5.471 trillion, down 17.2% on the same period of 2025 (Interfax, 9 October 2026), a smaller fall than the 38.3% recorded for the first four months. Natural decline, 2024 596,200 Up 20.4% year on year, before war losses. Births, Q1 2026 ~272,000 Down 6% — a two-hundred-year low. TFR 1.374. War dead, named and confirmed 223,539 3 July 2026. A hard floor; models run to ~352,000 and beyond. Oil and gas revenue −38.3% Year on year, first four months of 2026. What a second life is Most declining powers are somewhere on a single arc. Russia is on its second. It completed a full decline between 1989 and 1999 — imperial dissolution, industrial collapse, a mortality crisis with no peacetime precedent in an industrialised country. Then it executed a genuine rebound: from 2000 to roughly 2013, hydrocarbon revenue funded real reconstruction of state capacity, real income growth, and the restoration of something that could plausibly be called great-power standing. That rebound is over, and the second decline is structurally worse than the first for one reason. A recovery requires an asset to recover with . In 1999 Russia still had the demographic inheritance of the Soviet period — a large, educated, if damaged, working-age population. In 2026 it does not, and the war is consuming what remains at the precise age band that would otherwise have produced the next cohort. The first decline cost Russia its economy. The second is costing it the people who would have rebuilt one. The demographic arithmetic The last clean count was 146,028,325 on 1 January 2025. Natural decline in 2024 was 596,200 — a 20.4% deterioration in a single year. First-quarter 2026 births came in around 272,000, down 6%, which is the lowest rate in roughly two centuries. The total fertility rate stands at 1.374. The war losses sit on top of that, and the honest way to present them is as a range with the basis stated. Mediazona’s named-and-confirmed count reached 223,539 (including 7,345 officers) by 3 July 2026 — that is a floor , not an estimate, because it requires a name. The probate-based model puts deaths among men aged 18–59 at roughly 352,000 by end-2025. GCHQ’s May 2026 assessment is around 500,000. The regional concentration is extreme and politically significant: Tyva has lost 476 per 100,000, a rate that would be nationally destabilising if it were distributed evenly. It is not distributed evenly, which is precisely the point. The fiscal arithmetic The federal deficit ran 5.8 trillion roubles (~$81.8bn) in the first four months of 2026 alone, blowing past the annual target in a quarter. Oil and gas revenue — the engine of the entire second life — fell 38.3% year on year over the same period. A war economy can run on deficit spending for a long time. It cannot run on a collapsing revenue base indefinitely, and the sanctions architecture has removed most of the instruments that would normally bridge the gap. Update, 12 October 2026. The fall in energy revenue narrowed as the year went on. By January–September, oil and gas revenue was RUB 5.471 trillion, down 17.2% year on year, and the deficit was RUB 5.186 trillion (2.2% of GDP), with the full-year forecast raised to RUB 7.345 trillion (3.2% of GDP), according to preliminary Finance Ministry data reported by Interfax on 9 October 2026. The four-month figure above is left as published. The revenue base is weaker than in 2025 but the pace of the fall has eased, so the −38.3% is best read as a trough in the rate of change rather than a trend. The imperial ledger This is where Gilpin’s framework bites hardest, because the commitments are being shed , visibly, in every direction at once: Armenia’s arms dependence on Russia has collapsed from 96% to under 10%. It froze CSTO participation in February 2024, stopped paying dues in May 2024, and adopted an EU-accession bill in March 2026. The Africa Corps evacuated roughly 400 personnel from Kidal in April 2026 — the security-for-concessions model failing in its flagship theatre. In Central Asia, China has committed more than $35bn of FDI over the decade against Russia’s roughly $20bn, with trade at about $106.3bn. Russia is being quietly foreclosed on by its own strategic partner. The Aktau crash — 38 dead, an eventual admission of responsibility in October 2025, a compensation settlement in April 2026 — demonstrated that a Central Asian state could now extract an apology from Moscow. That is a measurement of relative power, not of diplomacy. What Russia retains is leverage rather than power: remittance dependence running as high as 40% of GDP in Tajikistan and Kyrgyzstan, roughly $11.5bn of Uzbekistan’s ~$15bn in remittances, and about 7,000 troops in Tajikistan. Leverage is what you hold when influence has gone. The succession accelerant Putin is 73, has held power for about 27 years, and would pass Stalin’s tenure in July 2030. There is no institutionalised succession mechanism, and the death of Ivanov in late June 2026 produced a visible jolt through the system precisely because there is no mechanism to absorb one. A personalist regime with a demographic hole, a war economy, a collapsing revenue base and no succession [continues on the page]

Europe: the continent that stopped compounding

Source: https://worldwar3.army/europe.html

Demographically exhausted, structurally dependent on migration it has not politically consented to, and losing the industrial base independently of both.

Chapter 04 · First decline, high dispersion The continent that stopped compounding. Three failures are running at once and they are usually discussed as one. Europe is demographically exhausted. It is structurally dependent on migration it has not politically consented to. And its industrial base is failing for reasons that have nothing to do with either. GERMANY SHOWN · SEE DISPERSION BELOW The short answer · updated 24 October 2026 Europe is in relative decline for three separate reasons: a spent demographic engine, a migration dependency it has not consented to, and an industrial base failing on energy costs. On the index the continent is not one trajectory: Poland is rising (+3.1) while Germany (−3.8), Italy (−4.0), France (−2.1) and the UK (−1.8) fall. EU workers per retiree fall from 2.7 to 1.5 between 2022 and 2050 (OECD). Irregular EU crossings fell about 40% year on year to roughly 39,000 in January to May 2026, but Western Mediterranean detections rose 46% (Frontex). German industrial output is about 15% below its 2017-18 peak (ifo / CE Interim, 2026). The US tariff on EU cars rose from 15% to 25% in July 2026 (Fortune). EU workers per retiree 2.7 → 1.5 2022 to 2050. Old-age dependency 36% → 55%. German industrial output −15% Against the 2017–18 peak. Energy-intensive output −15.2% since Feb 2022. Industrial electricity €0.25 Per kWh in Germany, against $0.08 in the US and €0.10 in China. AfD national polling 27–28% Early 2026 — a record, and ahead of the CDU. One: the demographic engine is spent This is the slowest and most certain of the three, and it is not really contested by anyone. EU old-age dependency moves from 36% to 55% and workers per retiree from 2.7 to 1.5 by 2050. Italy’s working-age population shrinks by more than a third by 2060. Germany is legislating retirement at around 68 by 2051 and roughly 70 by 2091, indexed to life expectancy; the Netherlands is on a similar path to about 70 by the 2060s; Italy restores its life-expectancy link in 2027. Raising the retirement age is not a solution. It is an accounting response to the absence of one. The only demonstrated escape from a fertility trap at this depth is inward migration at scale, and that is where the second failure begins. Two: the dependency Europe has not consented to Germany’s skilled-migration pivot is the honest version of the continent’s position: a country that has understood it cannot fund its pension system from its own birth cohorts and has built a two-door policy to import the difference. Every European state faces the same equation. Most have not said so out loud. Meanwhile the enforcement apparatus is expanding fast. Irregular EU crossings fell about 40% year on year in the first five months of 2026 to roughly 39,000, following a 26% fall across 2025. But the flow displaced rather than stopped: Western Mediterranean detections rose 46% over the same five months and 66% in the first quarter, and Ceuta land crossings rose several hundred per cent even as Canary Islands arrivals fell 78–83%. Mediterranean deaths rose 27% to 1,570 in the first half of 2026, with Central Mediterranean deaths up more than 150%. The institutional response is hardening in parallel: the Return Regulation agreed in June 2026 permits detention up to 24 months; the Entry/Exit System reached full rollout in April 2026; the UK is paying France £662m for enforcement across 2026–29; Germany has resumed Afghanistan deportation flights and extended family-reunification waits from 12 to 18 months. A continent that needs migrants to fund its pensions, is politically unable to accept them, and is building a detention-and-externalisation apparatus to manage the contradiction. That is the actual story, and it is bleak enough without embellishment. The political repricing is already visible. The AfD polls at 27–28% nationally, ahead of the CDU. That is not a protest number any more; it is a governing-coalition number, and every mainstream European party is now writing policy in its shadow. Three: the industrial base, failing independently This failure has nothing to do with the first two, which is why conflating them produces bad analysis. German industrial production sits roughly 15% below its 2017–18 peak. Energy-intensive production is down 15.2% since February 2022. Industrial electricity runs about €0.25/kWh against $0.08 in the United States and €0.10 in China — a cost gap no amount of engineering excellence closes. The consequences are arriving as headcount. Roughly 124,000 industrial jobs went in 2025, about 50,000 of them in autos. Bosch is cutting around 22,000; ThyssenKrupp around 11,000 with steel capacity going from 11.5Mt to under 9Mt; the chemical sector is running at about 70% capacity, a historic low. Volkswagen has proposed up to 100,000 cuts — roughly 15% of its workforce — with capacity falling from about 12 million vehicles pre-pandemic to a target near 9 million, models halved and variants cut by three-quarters. Q1 2026 net profit fell 28% to €1.56bn. BASF is putting €8.7bn into Zhanjiang while cutting 2,800 at Ludwigshafen. The US tariff on EU cars went from 15% to 25% in July 2026, which ifo and Kiel price at roughly €15bn of output in the first year. Germany’s answer is to become a different kind of economy: defence spending from €108.2bn toward about €200bn by 2030, approaching a third of the federal budget, with a €500bn infrastructure fund that had deployed only about €11bn of its first €40bn tranche by mid-2026. Whether that is reindustrialisation or a war economy depends on which decade you ask. The dispersion is the story Treating “Europe” as one trajectory is the most common analytical error in this subject. Within the index, Poland is rising at +3.1 — Europe’s fastest-rearming state and one of its fastest-growing large economies. Germany is falling at −3.8. Italy is at −4.0 and is the only nation in the index placed in terminal decline, because its working-age contraction has no recovery instrument attached to it. France is at −2.1 and is the clearest live case of Gilpin’s ledger — its African base network has collapsed, and Russian and Turkish arrangements moved into the space it vacated. The United Kingdom is at −1.8, with financial centrality outliving the productive base that created it: it is now the second-largest foreign holder of US Treasuries at $897bn, which is a statement about the City rather than about Britain. The evidence Figure Value Date Source Confidence EU old-age dependency 36% → 55% 2022→2050 OECD Med–High EU workers per retiree 2.7 → 1.5 2022→2100 OECD Med–High Italy working-age population −>1/3 by 2060 2060 OECD Med–High German retirement age ~68 (2051) → ~70 (2091) 2031– Bruegel Medium EU irregular crossings ~39,000 (−40% YoY) Jan–May 2026 Frontex High Western Med detections ~7,100 (+46% YoY) Jan–May 2026 Frontex High Canary Islands arrivals −78% to −82.6% Q1 2026 Spain Interior Ministry Med–High Mediterranean deaths 1,570 (+27% YoY) H1 2026 IOM Medium UK payment to France £662m 2026/27–2028/29 UK Home Office High Return [continues on the page]

Southeast Asia and Australia — the quiet winners | The Relative Decline Project

Source: https://worldwar3.army/asia-pacific.html

Relative decline is a ratio. Somebody is on the other side of it — and Australia is the empirical answer to Europe's demographic problem.

Chapter 05 · Rising The quiet winners of somebody else’s decline. Relative decline is a ratio. If the incumbents are losing share, somebody is gaining it — and the gainers are mostly not the powers that get written about. This is the chapter that keeps the site honest, because a model in which everyone declines is a model that has measured nothing. INDONESIA SHOWN ASEAN-5 growth, 2026 4.2% Wider subregion 4.7% in 2026, 4.8% in 2027. Vietnam growth 6.3–7.6% The fastest in the region, on a tech upcycle and supply-chain diversion. Indonesian nickel output 2.6Mt Of roughly 3.9Mt worldwide. The Philippines is second at ~270,000t. Australian institutional score 78 The highest in the index. Demographic engine 68, against Italy’s 16. What winning looks like at this point in the cycle ASEAN-5 is growing at 4.2% in 2026, flat against 2025, with the wider subregion at 4.7% and 4.8% in 2027. Vietnam leads at 6.3–7.6%, the Philippines at 4.4–5.1%, Indonesia at 5.0–5.1%. These are not spectacular numbers by the standards of the Asian miracle. They are spectacular by the standards of 2026, when the four largest incumbent economies in this index are all scoring negative velocity. The structural tailwinds are the ones the incumbents have lost: demographic dividends still running, deepening regional integration, supply-chain diversification arriving as capital, and a services and digital sector growing from a low base. Bain, DBS and Vriens describe the coming decade as one of divergence within the region rather than uniform ascent, which is the right frame — Vietnam and Indonesia are not on the same path as Thailand or Myanmar. Indonesia is the headline case Roughly 280 million people, BRICS membership from January 2025, a stated ambition to join the OECD within about three years, and control of roughly 2.6 million tonnes of the world’s 3.9 million tonnes of nickel. Mining is about 9% of GDP and 17% of exports. The Danantara sovereign fund is reported around $900bn in target assets — a figure worth treating cautiously, but the intent is unambiguous. The growth target is 8% by 2029, up from about 5%. What makes Indonesia genuinely interesting is not the growth. It is the hedging . It institutionalised a 2+2 dialogue with China in April 2025. It became the third BrahMos customer after the Philippines and Vietnam, and signed more than a dozen agreements with India in July 2026. It signed 26 agreements with Singapore in the same month. It pledged roughly 8,000 troops to the Gaza stabilisation force. This is a middle power learning to be courted by everyone and captured by no one, and it is doing it well. The vulnerability is the commodity: nickel at roughly $30,000 a tonne in 2022 is projected at $14,000–15,000 in 2026. A resource-financed industrial strategy is a bet on a price. Vietnam, and the clock that is already running Vietnam has executed bamboo diplomacy about as well as it can be executed: upgraded relations with the United States, China, Japan and Korea near-simultaneously, absorbed the largest share of China-plus-one manufacturing relocation, and kept a maritime dispute with Beijing from becoming a strategic rupture. The caution this chapter should carry: Vietnam’s own fertility rate is falling, and its demographic window closes well before its income level catches up. It has perhaps two decades to convert a manufacturing position into a technological one. Korea managed that transition. Thailand did not. Australia — the empirical answer to Europe Australia does not belong in this chapter for its growth rate. It belongs here as the policy case, and it is arguably the most important single data point in the whole index. Europe, Japan, Korea, Italy and China all face the same fertility trap. Australia has faced it too — its own birth rate is below replacement and has been for decades. The difference is that Australia has run a large, selective, skills-weighted migration programme for thirty years and has therefore simply not had the demographic cliff. It scores 68 on the demographic engine against Italy’s 16 and Korea’s 8, and 78 on institutional capacity, the highest reading in the index. The lesson is not “migration is good.” It is narrower and more useful: chosen migration, at scale, sustained across decades, with public consent maintained through visible selectivity, is the only demonstrated escape from a fertility trap. The honest caveat has to be carried on the same page, because it is what makes the lesson real rather than smug. Australia is paying for this in housing, and the consent is fraying under that pressure. Canada is the warning case: a comparable programme, run faster, without matching housing supply, produced a political reversal in 2025–26 that has put the whole model in question. Selectivity buys consent. Volume without absorption capacity spends it. Japan and Korea — the counter-examples in the same region Both sit in this chapter for contrast rather than company. Japan runs a TFR of 1.14 with natural decline above 900,000 in 2025, roughly 30% of the population over 65, and an old-age dependency ratio moving from about 54.9 to 80.7 per 100 by 2050. Its response has been the most competent management of the worst demography anywhere — a foreign-worker target of roughly 1.23 million through FY2028, foreign residents up 9.6% to about 4 million. It is still short: about 770,000 of the 4.19 million workers it needs by 2030. Korea is worse and better at once: a TFR that bottomed at 0.72 and has recovered to roughly 0.80 — still the lowest on earth — deaths exceeding births by 108,900 in 2025, and roughly 360 trillion won spent cumulatively on pro-natal policy with very little to show for it. Korea scores 76 on the productive engine and 8 on the demographic one. That combination is the clearest demonstration in the index that a world-class economy cannot outrun an absent generation. The evidence Figure Value Date Source Confidence ASEAN-5 growth 4.2% 2026 IMF / Krungsri Medium (forecast) Developing SE Asia growth 4.7% / 4.8% 2026 / 2027 ADB Outlook Medium (forecast) Vietnam growth 6.3–7.6% 2026 ADB / McKinsey Medium Indonesia population ~280 million 2025 East-West Center High Indonesia BRICS membership Joined 6 Jan 2025 SCMP High Indonesian nickel output ~2.6Mt of ~3.9Mt world 2025 ISEAS / Fulcrum High Nickel price ~$30k/t → ~$14–15k 2022→2026 ISEAS Medium Indonesia growth target 8% by 2029 2026 Fortune / EAF Medium (target) Danantara sovereign fund ~$900bn reported 2025–26 Fortune / The Diplomat Low–Medium Singapore–Indonesia agreements 26 6 Jul 2026 PMO Singapore High Japan TFR 1.14 (record low) 2025 Nippon.com High Japan natural decline >900,000 2025 Japan Times High Japan old-age dependency 54.9 → 80.7 per 100 2026→2050 OECD Med–High Japan foreign-worker target ~1.23 million through FY2028 Japan Times High Japan 2030 worker shortfall 770,000 of 4.19M needed by 2030 AMRO Medium South Korea TFR 0.72 → ~0.80 2023→2026 CNN High Korea deaths minus births 108,900 2025 CNN High Korea pro-natal spending ~360tn won (~$250bn) cumulative CNN Medium The five engines (Australia) [continues on the page]

The new frontiers — Greenland, the Arctic, the sea routes | The Relative Decline Project

Source: https://worldwar3.army/frontiers.html

When incumbent powers cannot grow their share of a fixed world, they go looking for unclaimed world. This is where decline turns into acquisition.

Chapter 06 · Acquisition The new colonial frontiers. When powers cannot grow their share of a fixed world, they go looking for unclaimed world. Greenland, the Arctic seabed, the Antarctic minerals moratorium and the sea routes are where relative decline stops being an economic story and becomes a territorial one — and it is happening through licences and export credit long before it happens through force. FLASHPOINT: ARCTIC / GREENLAND · 2 OF 5 CONDITIONS MET EU critical raw materials present in Greenland 25 of 34 2023 survey. Eighth globally for rare-earth reserves at ~1.5Mt. Greenland rare earths commercially mined 0t As of 2026. The entire contest is over a resource nobody has extracted. Arctic winter sea-ice maximum 14.29M km² 15 March 2026 — tied record low. JAXA recorded an outright low of 13.76M. First US Arctic cutter delivery 2028 Up to 11 ships, $6.8bn+. The gap between now and then is itself a window. Greenland: a contest over nothing yet extracted A 2023 survey found 25 of the 34 minerals the European Commission classifies as critical raw materials present in Greenland. It ranks eighth globally for rare-earth reserves at roughly 1.5 million tonnes, and holds two of the largest known deposits anywhere in Kvanefjeld and Tanbreez, alongside copper, graphite, gallium, tungsten, zinc, gold, silver, iron ore, platinum, molybdenum, tantalum and vanadium. The commercially mined total to date is zero tonnes. That is the most important number on this page, and it is why the Greenland story is about exclusion rather than extraction. The US Export-Import Bank’s June 2025 letter of interest — $120 million toward Tanbreez, which would be the administration’s first overseas mining investment — is small money for a strategic option. The framing around the broader arrangement has been explicit: the value is the power to determine who may not mine there, and specifically to exclude Chinese and Russian capital and any military footprint. Washington is not buying rare earths. It is buying a veto, well in advance of anyone needing the ore. The Arctic realities argue for patience: the deposits are remote, the ore bodies are complex, some carry uranium co-products with their own politics, Greenlandic consent is a live and legitimate constraint, and permitting-to-production timelines run a decade or more. This is a twenty-year play being conducted at the tempo of a five-year news cycle, which is usually how acquisition phases begin. The Arctic: a claim dressed as a shipping route The ice is going. The March 2026 winter maximum came in at 14.29 million km² on the NSIDC series, tied for the lowest on record, and JAXA recorded an outright record low of 13.76 million km² two days earlier. What opens is not a trade route in any commercial sense yet — Northern Sea Route cargo reached 37.02 million tonnes in 2025, down 2.3%, missing Moscow’s 80-million-tonne target for the second consecutive year. But China-linked NSR container transits rose to 14–15 in 2025 from 11 in 2024 and 7 in 2023, with tonnage up roughly 2.6x to about 400,000 tonnes. That is a small number growing fast, and it is being used to establish presence rather than to move cargo economically. The militarisation is unambiguous and is being funded on both sides. Russia has extended the Nagurskoye runway from about 2,500m to 3,500m. Denmark has committed DKK 27.4bn (~$4.26bn) in a second Arctic agreement on top of DKK 14.6bn in January 2025. The United States has contracted up to 11 Arctic Security Cutters for more than $6.8bn — five ships to Davie in May 2026, four to Bollinger and two to Rauma in Finland in July — with first delivery in 2028 and the six-ship tranche complete in 2031. A power that will not have the hulls until 2028 is a power operating inside a known capability gap until then, and known gaps are precisely what this index means by a window. The Central Arctic Ocean fishing ban covers about 2.8 million km² through at least 2037. It is the only part of the Arctic where the powers have agreed in advance not to compete — and it covers fish, not minerals. The sea routes: where a declining power imposes cost cheaply The strategic logic of chokepoints is that they invert the usual relationship between capability and leverage. A power in decline cannot build a navy to match a rising one, but it can make a strait expensive, and expensive is often enough. Hormuz. The clearest live demonstration in this index. The 2026 conflict spiked oil, battered Asian currencies, and pushed Japan and China into selling roughly $88bn of US Treasuries in a single month to fund currency defence. A regional resource conflict reached into the hegemon’s funding market inside four weeks. Tanker war-risk premia are the best real-time indicator on this site. Malacca. The structural vulnerability of every economy in the Asia-Pacific chapter, and the reason Kra Canal and land-bridge proposals never quite die despite never quite penciling. The Danish Straits. Where the shadow-fleet question meets the NATO-lake question, and where an interdiction decision has Article 5 implications that nobody wants to test. The Turkish Straits. Montreux gives Ankara a wartime instrument of real value, which is a meaningful part of why Turkey scores as holding rather than slipping despite a fiscal engine at 29. Subsea cables. The cheapest asymmetric target in existence, in the Arctic and the Baltic alike, and the one where attribution is hardest and therefore where deterrence is weakest. Antarctica: the deadline nobody is discussing The Antarctic Treaty’s Environmental Protocol prohibits mineral resource activity, and the earliest point at which any party can request a review conference is 2048. That is inside the planning horizon of every state in this index. Chinese and Russian station construction with dual-use characteristics, and the deadlock in the CCAMLR krill negotiations, are both best read as positioning for a conversation that formally cannot begin for two decades — which is exactly how long a serious acquisition strategy takes. The evidence Figure Value Date Source Confidence EU critical raw materials in Greenland 25 of 34 2023 survey European Commission-linked Medium Greenland rare-earth reserves ~1.5Mt (8th globally) USGS-linked CSIS / Yale E360 Low (estimate) Greenland rare earths mined 0 tonnes as of 2026 Yale E360 / CSIS High US EXIM letter of interest, Tanbreez $120m (letter, not closed) Jun 2025 CSIS / CNBC Medium Arctic sea-ice max (NSIDC) 14.29M km² (tied record low) 15 Mar 2026 NASA / NSIDC High Arctic sea-ice max (JAXA) 13.76M km² (record low) 13 Mar 2026 JAXA / NIPR High NSR total cargo 37.02Mt (−2.3% YoY) 2025 Moscow Times / gCaptain High Moscow’s NSR target 80Mt (missed 2nd year) standing goal multiple Medium (target) China-linked NSR transits 14–15 (vs 11 in 2024, 7 in 2023) 2025 gCaptain / High North News Medium China-linked Arctic tonnage ~400,000t (~2.6x YoY) 2025 High North News / CHNL Medium Nagurskoye runway extension ~2,500m → 3,500m through 2025–26 CSIS “Ice Curtain” High Denmark Arctic agreements [continues on the page]

Africa — improving, and being bought | The Relative Decline Project

Source: https://worldwar3.army/africa.html

Africa's condition is genuinely improving and its resource base is being acquired on terms it does not set. Both are true. The question is value addition.

Chapter 07 · Rising base, contested ownership Improving, and being bought. Africa’s condition is genuinely improving and its resource base is being acquired on terms it does not set. Both statements are true at the same time, and most writing on the subject collapses them into one — either a rising-continent story or a new-scramble story. The interesting analysis lives in the gap between them. NIGERIA SHOWN Nigeria demographic engine 82 The highest reading in the index — attached to an institutional score of 26. Africa Corps evacuated, Kidal ~400 April 2026. Security-for-concessions failing in its flagship theatre. One-sided violence, 2025 ~76,500 Deaths worldwide, up over 400%, driven principally by Sudan. Highest since 1994. South Africa velocity −3.4 Outsized diplomacy on a shrinking material base. The improvement is real It is worth stating plainly before the rest of the chapter, because the resource story tends to swallow it. Electrification is advancing. Mobile money has leapfrogged a banking system that was never built, and in several markets the payments infrastructure is ahead of Europe’s. The African Continental Free Trade Area exists and is slowly becoming operational. Urbanisation is producing genuine megacity economies. And the youth bulge is, in pure arithmetic terms, the largest demographic asset on the planet at a moment when every incumbent power in this index is running out of workers. Nigeria scores 82 on the demographic engine — the highest reading anywhere in the model, above India’s 74 and far above the United States at 62. That is what a dividend looks like on paper. The acquisition is also real The same continent is the object of the most concentrated resource competition since the 1880s, and the instruments have changed while the structure has not. China runs resource-for-infrastructure lending across the copper belt, the cobalt and lithium provinces of the DRC, Zambia and Zimbabwe, and has moved decisively into the battery-mineral chain upstream of its own dominant refining position. Gulf sovereign wealth has moved in behind, buying ports, logistics and agricultural land with a longer horizon and less conditionality than either Beijing or Brussels. Russia trades security for mining concessions through the Africa Corps — a model now visibly failing, with roughly 400 personnel evacuated from Kidal in April 2026, but one that established a template others are copying. Turkey trades drones for influence, and has built real positions in the Sahel on a fraction of anyone else’s budget. France has lost its base network almost entirely, which is the single clearest entry in Gilpin’s ledger anywhere in this index — and the vacuum it left is what the four above moved into. The question that actually decides it: value addition Whether Africa rises or is merely extracted from turns almost entirely on one variable, and it is not geology. It is whether critical minerals are refined on the continent or shipped raw. A country that exports ore imports poverty with a delay. A country that exports refined product imports an industrial base. The deposits are identical. The outcome is not. Indonesia is the template everyone is watching: a raw-ore export ban that forced smelting onshore and converted a mining sector into an industrial one, at real cost and with real friction, but successfully enough that it now anchors a middle power’s entire growth strategy. The DRC and Zimbabwe are the live African tests of the same logic, and both are attempting it with weaker institutions and more leverage against them. That makes this a governance question, and it is where the index is pessimistic. Nigeria pairs the best demographic engine in the model with an institutional score of 26 and a cohesion score of 28. South Africa runs a productive engine of 26 and a velocity of −3.4 — the clearest case anywhere of diplomatic influence outliving material capacity, conducting great-power diplomacy on a shrinking base. The continental picture on governance is not encouraging either: Freedom House recorded a twentieth consecutive year of global freedom decline with 54 countries deteriorating against 35 improving, and the largest single-country decline in the 2026 cycle was Guinea-Bissau at −8. The cost of the vacuum, stated honestly The Sahel and Horn are already the largest single source of organised-violence fatalities in the world. One-sided violence — deliberate killing of civilians — rose more than 400% in 2025 to roughly 76,500 deaths, the highest level since the Rwandan genocide in 1994, driven principally by Sudan. Total fatalities from organised violence reached 244,600 against about 187,000 the previous year. This is not a future risk on the flashpoint board. It is the present cost of a security vacuum that opened when the French drawdown met a set of resource-for-protection arrangements that could extract but could not stabilise. Any honest accounting of the “new scramble” has to carry this number, because it is what the scramble has cost so far. The evidence Figure Value Date Source Confidence Organised-violence deaths 244,600 (from ~187,000) 2025 UCDP / Uppsala High One-sided violence deaths ~76,500 (+400%+) 2025 UCDP High State-based conflicts worldwide 65 (record since 1946) 2025 UCDP High Africa Corps evacuated, Kidal ~400 personnel 26 Apr 2026 CNN High Russian personnel in Mali ~2,500 early 2026 Al Jazeera Medium Years of global freedom decline 20 consecutive through 2025 Freedom House High Countries declining vs improving 54 vs 35 2026 cycle Freedom House High Largest single-country decline Guinea-Bissau, −8 2026 report Freedom House High Countries actively autocratising 44 (41% of world population) 2026 report V-Dem High Autocracies vs democracies 92 vs 87 2026 report V-Dem High Nigeria demographic engine score 82 (highest in index) 2026 this index Medium Nigeria institutional score 26 2026 this index Medium South Africa velocity −3.4 2026 this index Medium The five engines (Nigeria) The most extreme profile in the index, and the one that best illustrates why a single decline score would be useless. Nigeria has the strongest demographic engine in the model and among the weakest of everything else. The steelman The “new colonialism” framing patronises African states by denying them agency they demonstrably have. Chinese lending terms have been renegotiated repeatedly and often downward. Several governments have cancelled, repriced or nationalised projects. Indonesia’s export-ban template is being studied and adapted precisely because African governments understand the value-addition problem perfectly well — they are not unaware of it, they are negotiating from a weak position, which is a different thing. And the competition itself has value. A continent with four or five suitors gets better terms than one with a single creditor, which is exactly the position much of Africa was in under structural adjustment. The Gulf’s arrival, Turkey’s arrival and India’s growing interest are improvements in bargaining power, not additional predations. The improvement metrics — electrification, payments, trade integration, child mortality — are real, measurable, and have continued through the entire period the scramble narrative covers. What would falsify this chapter Refining capacity for cobalt, lithium or copper reaching a meaningful share of continental output; AfCFTA intra-regional trade passing 25% of [continues on the page]

South America — the middle-quality institutional trap | The Relative Decline Project

Source: https://worldwar3.army/south-america.html

Living standards rose. Governance did not keep pace. The growth ceiling in Latin America is institutional, not geological.

Chapter 08 · Plateau The middle-quality institutional trap. The academic term is exact and worth adopting: flawed democracies paired with low-to-medium capacity states. Living standards have risen across the region for two generations. Governance has not kept pace, and the growth ceiling that results is institutional rather than geological. BRAZIL SHOWN Brazil institutional engine 40 Against a productive engine of 38 and a demographic engine of 52. Mexico institutional engine 36 The nearshoring dividend running against cartel-driven erosion. Essequibo conditions met 2 of 5 The purest resource case on the flashpoint board, and the least watched. Brazil velocity −0.4 Effectively flat. Not declining — simply not compounding. The divide that matters, and it is not left versus right The single most useful finding in recent Latin American analysis is that the 2026 performance split does not track ideology at all. It tracks institutional quality. Countries with strong institutions and modest growth (Chile, Uruguay) have consistently outperformed countries with strong growth and weak institutions (Brazil, Mexico.) The real divide is between states that can sustain credible, rules-based environments and those that continue to rely on discretionary, personality-driven politics — and governments of every political colour appear on both sides of it. That finding is inconvenient for almost everyone writing about the region, which is a reasonable proxy for it being true. Why the ceiling is institutional Latin America has the resource endowment, the demographic profile and the market access that East Asia did not have when it began its ascent. It has not converted them, and the conventional explanations — commodity dependence, external interference, inequality — are symptoms rather than causes. The structural account is better: the tension between an industrialisation path and primitive accumulation, combined with institutional deficiencies in technology and industrial policy, produced a region that could grow but could not upgrade . Middle-income countries must build strong institutions at precisely the moment societal demand for those institutions is weakest — the population is no longer poor enough to demand basic state capacity and not yet rich enough to demand sophisticated regulation. That gap is where the trap lives, and it is a political problem rather than an economic one. Chile and Uruguay are not growing faster than Brazil. They are compounding more reliably, and over thirty years reliability beats velocity. The 2026 pressures compound it. Fiscal space is squeezed, debt burdens are elevated after a decade of low growth, and the external environment has turned: rising protectionism, intensifying great-power rivalry over industrial and technological supremacy, and the visible withdrawal of major powers from multilateral cooperation. A region whose growth model depends on open markets and rules-based arbitration is losing both at once. Argentina: the live experiment Argentina is worth its own section because it is testing a genuine hypothesis: whether an institutional reset can be imposed from outside the institutions, by a government that treats the existing apparatus as the problem rather than the instrument. The dollarisation endgame and the Milei austerity programme are, whatever one thinks of them, the most consequential natural experiment in the region in decades. If it works, the middle-quality trap has an escape route that does not require the slow institutional accumulation Chile took forty years to build. If it fails, the failure will be read as confirmation that the trap is genuinely structural. This index takes no position on the outcome and will score it when the data resolves. Mexico: growth and erosion in the same column Mexico is the clearest case of the divide running through a single country. The nearshoring dividend is real — proximity to the United States at the exact moment supply chains are being re-shored is the best structural hand in the region. Against it runs cartel-driven institutional erosion that shows up in an institutional score of 36 and a cohesion score of 35. Mexico is simultaneously the region’s best-positioned economy and one of its weakest states, and the index scores it at +0.3 because those two facts almost exactly cancel. Essequibo: the resource case nobody is watching Venezuela’s claim on Guyana’s Essequibo region is the purest test of this site’s central thesis anywhere on the flashpoint board, and it attracts a fraction of the attention that Taiwan or the Baltic does. Every element is present: a collapsing petro-state with acute legitimacy pressure; an adjacent neighbour sitting on one of the largest offshore oil discoveries of the century; a territorial claim already formalised domestically; a vastly asymmetric military balance; and an international legal process whose authority the claimant has reason to reject. It meets two of the five conditions formally, but the fixed resource condition is met about as purely as it can be, and the legitimacy-pressure condition is severe. This index puts it at 10–18% over five years and regards that as the most under-priced number on the board. The evidence Figure Value Date Source Confidence Institutional trap characterisation Flawed democracy + low/medium state capacity 2024–26 Cambridge UP High (academic) 2026 performance split Strong institutions > strong growth 2026 regional analysis Medium Brazil velocity −0.4 2026 this index Medium Brazil institutional / productive 40 / 38 2026 this index Medium Mexico velocity +0.3 2026 this index Medium Mexico institutional / cohesion 36 / 35 2026 this index Medium Essequibo conflict band (5yr) 10–18% 2026–31 this index Low (subjective) Countries actively autocratising 44 (41% of world population) 2026 V-Dem High Global democracy level Back to 1978 baseline 2026 V-Dem High The five engines (Brazil) The steelman South America is the safest large region on earth, and that is worth more than growth. It has no nuclear-armed rivalries, no great-power land border, no serious interstate war in eighty years outside the Essequibo file, and a functioning nuclear non-proliferation arrangement between its two largest states that the rest of the world failed to replicate. It sits outside every quadrant of the great-power contest this site describes. In a decade where the index expects conditions for conflict to worsen almost everywhere, that is not a small asset. Substantively: Brazil has genuine strategic autonomy and real Global South convening power; Pix demonstrated payment sovereignty in a way the G7 has not matched; Chile’s lithium state-partnership model is a serious attempt at value addition; and the region holds the resource endowment that the energy transition requires. “Messy” is a description of the politics, not a forecast of the outcome. What would falsify this chapter Brazil or Mexico showing sustained institutional-quality improvement on independent indices for five consecutive years; Argentina’s programme delivering both inflation stabilisation and institutional durability past a change of government; or regional intra-trade and technology exports rising fast enough to break the middle-technology pattern. What to watch Argentina past the next electoral cycle. Reforms that do not survive a government change were never institutional. Venezuelan deployments east of the Cuyuni and any licensing round inside the disputed block. Chile’s lithium partnership model — the region’s best value-addition attempt, and a direct comparison with Indonesia’s. Mexican institutional indices against nearshoring inflows. [continues on the page]

About, standards and corrections | The Relative Decline Project

Source: https://worldwar3.army/about.html

What this site is, what it refuses to do, the editorial standards it holds itself to, and the corrections log.

Standards What this is, and what it refuses to do. The Relative Decline Project measures how fast the major powers are losing or gaining relative position, and prices the conditions under which that contest turns violent. It lives at worldwar3.army, which is a deliberate choice and a demanding one. The domain A domain that promises apocalypse and delivers charts reads as clickbait. A domain that promises apocalypse and delivers rigour reads as authority. The distance between those two outcomes is the entire editorial strategy of this site, and two rules close it. The site prices conflict. It does not predict it. Every claim about fighting is a conditional with named triggers and a published probability band. “Here is what would have to be true” is a statement that survives contact with 2027. “Here is what will happen” is not. The masthead is the argument, not the domain. This is the Relative Decline Project. worldwar3.army is where it lives. The URL carries the register; the content has to carry the weight. Editorial standards These are published here so readers can hold the site to them. Every figure carries a date, a source and a confidence grade. No exceptions, including for figures that are obviously true. Every chapter carries a steelman of the opposing case, written to persuade rather than to be knocked down. If the counter-case on a page reads weakly, that is a defect in the page. Every forecast carries a resolution date and is graded in public, including the misses, with the original text unedited. No prophecy. Conditionals and probability bands only. Single-source claims are labelled as such and are never presented alongside multiply-confirmed ones without the distinction being visible. Ranges stay ranges. Where credible estimates differ by a factor of two, all of them are published with their bases stated. Corrections are logged, dated and left visible. Nothing is quietly edited. Every weight is published. If you disagree with a score, the method page tells you exactly which number to argue with. What this site refuses to do It does not treat decline as moral failure. A low engine score is a measurement. Europe trading global weight for stability and longevity may be a rational choice rather than a defeat. It does not forecast collapse. Every nation in this index will still exist in 2050. Relative decline concerns share, and share is a ratio. It does not use civilisational framing. The Europe chapter engages Huntington’s thesis directly and at length, presents the evidence for it, and explains why the site does not adopt it — principally that intra-civilisational conflict massively outnumbers inter-civilisational conflict in every dataset, that civilisations are not actors capable of forming intent, and that a framework which cannot explain 1914 and 1939 is not a framework for explaining Europe. The migration pressure and its political consequences are real and are covered in detail. The frame is what is rejected, not the evidence. It does not collapse the engines into a single score. One number would be more shareable and would measure nothing. Sources The index is built on official statistics wherever they exist: IMF Article IV consultations and the World Economic Outlook, the Congressional Budget Office, Rosstat, China’s National Bureau of Statistics, Destatis and ifo, Japan’s e-Stat, OECD projections, Frontex, the Uppsala Conflict Data Program, Freedom House, V-Dem, SIPRI, NSIDC and JAXA, the USGS, and national treasuries and defence ministries. Where official data is contested or absent — Russian casualties being the clearest case — independent counts are published alongside their methodologies rather than in place of them. The underlying research base is a corpus of sourced country and issue files maintained continuously, each carrying the same figure/value/date/source/confidence structure used on these pages. Corrections log No corrections issued Corrections will be listed here with the date issued, the original claim, the corrected claim and the evidence that prompted the change. An empty log this early is expected, not reassuring. Revision schedule Version Scored Status Changes v1.0 22 Sep 2026 Current Initial publication — 20 nations, 8 flashpoints v1.1 31 Dec 2026 Scheduled Quarterly re-score; first forecast resolutions How to disagree with this site usefully Argue with an indicator, not with the conclusion. Every score on this site decomposes into published inputs. If you think China’s productive engine should be 85 rather than 71, the robot-density and labour-productivity series are the place to make that case — and if the evidence supports it, the score changes and the change is logged. That is the whole point of publishing the weights. The single most valuable thing a reader can send is a figure that contradicts one on these pages, with its source attached. Continue The method and weights · The five engines · Back to the index

Russia Defence Spending 2027 and Budget Deficit

Source: https://worldwar3.army/reports/velocity-2026-W44.html

Russia plans 17.1tn roubles of defence spending in 2027, 27% above plan, as the central bank holds at 14% and lifts its inflation forecast to 7.1-7.6%.

Velocity Report 2026-W44 Russia's war budget meets a 14% interest rate. Russia's draft budget puts 2027 defence spending at 17.1 trillion roubles, about 27% above the first plan, while the Bank of Russia held its key rate at 14% on 23 October and raised its 2026 inflation forecast to 7.1–7.6%. Spending, interest rates and falling oil and gas revenue now compete for the same fiscal room. That is a condition, not a forecast of collapse, and it bears on one number in the index: Russia's fiscal engine. Published 26 OCT 2026 · Updated 26 OCT 2026 9 min 2027 defence spending (draft) 17.1tn ₽ About $202.6bn; ~27% above the 13.5tn first budgeted (Reuters, 28 Sep 2026). 2026 deficit estimate 3.2% GDP Up from 1.6% in the budget law (Reuters, 28 Sep 2026). Key rate, 23 Oct 2026 14% Held. Annual inflation 6.8% on 19 Oct (TASS, 23 Oct 2026). 2026 inflation forecast 7.1–7.6% Raised from 6–7% (Moscow Times, 23 Oct 2026). Short answer. Russia plans to spend 17.1 trillion roubles on defence in 2027, the highest figure since the invasion of Ukraine in 2022, and the central bank held its rate at 14% on 23 October because inflation is running above its path. The two facts are linked: the central bank has said that if fiscal parameters change, monetary policy may need to change with them. On the Relative Decline Index, Russia's fiscal engine already scores 26 out of 100 and its decline velocity is −5.6; this week's data are consistent with that scoring, and no score has been changed. Key facts: sourced claims for citation Russia's draft budget documents show 17.1 trillion roubles (about $202.6 billion) of defence spending in 2027, roughly 27% above the 13.5 trillion roubles originally planned (Reuters, 28 September 2026). The 2026 deficit estimate in the draft budget rises to 3.2% of GDP from 1.6%, and 2026 net borrowing rises 26% to 5 trillion roubles (Reuters, 28 September 2026). The Bank of Russia held its key rate at 14% on 23 October 2026 and raised its 2026 inflation forecast to 7.1–7.6% from 6–7% (TASS and Moscow Times, 23 October 2026). Projected state debt reaches 21.7% of GDP in 2027, up from 19.9% in 2026 and above the 20% level authorities consider safe (Reuters, 28 September 2026). The 2026 oil and gas revenue estimate was cut to 7.6 trillion roubles from 8.9 trillion, a fall of about 15% (approx., calculated from Reuters, 28 September 2026). What moved this week: a hold that carries a fiscal message On Friday 23 October the Bank of Russia kept its key rate at 14%, as markets expected. The decision ended a run of ten consecutive cuts between June 2025 and July 2026, and followed a hold on 11 September. What changed was the forecast. The bank now expects 2026 inflation of 7.1–7.6%, against 6–7% in September. Annual inflation was 6.8% as of 19 October; seasonally adjusted price growth averaged 8% annualised in the third quarter against 4.8% in the second, and core inflation rose to 6.2% from 4.5% (TASS, 23 October 2026). The bank also said that if fiscal policy parameters change, monetary policy may need adjusting. That sentence matters more than the hold. It is the central bank telling the finance ministry that the two are drawing on the same pool of room. Its baseline puts the average 2027 key rate at 12.5–13.5% and 2026 GDP growth at 0.5–1% (Moscow Times, 23 October 2026). How much will Russia spend on defence in 2027? Draft budget documents seen by Reuters on 28 September put 2027 defence spending at 17.1 trillion roubles, about $202.6 billion at 84.4 roubles to the dollar. That is roughly 27% above the 13.5 trillion first budgeted for the year, and the highest figure since the war in Ukraine began. The three-year defence total in the documents is 50 trillion roubles. The 2026 figure is planned at 12.1 trillion roubles, but the actual outturn is classified and some military spending may sit in other budget lines (Reuters, 28 September 2026). For scale, 12.1 trillion roubles is about a quarter of the 48.6 trillion roubles of total 2026 spending (approx., calculated from the same documents). The step from 12.1 to 17.1 trillion is a rise of about 41% (approx.) across one budget year, though the 2026 figure is a plan, not an outturn, so the true increase is uncertain. We publish the range of reading, not one number. What do the deficit and debt numbers say? The 2026 spending plan rises 13.2% to 48.6 trillion roubles, or 20.9% of GDP. The deficit estimate rises to 3.2% of GDP from the 1.6% in the budget law. Net borrowing for 2026 rises 26% to 5 trillion roubles, and the government plans to draw 459 billion roubles from the National Wealth Fund. Total 2027 borrowing is projected to rise 43% to 7.7 trillion roubles (Reuters, 28 September 2026). This was visible earlier. In July the government’s budget portal implied a 2026 deficit of 4.83 trillion roubles against 3.79 trillion in the budget law, an overrun of about 1.04 trillion roubles, with first-half 2026 spending already producing a 5.73 trillion rouble deficit, 2.5% of GDP and 1.7 times the same period a year earlier. The finance ministry had also pushed back the date for restoring a zero primary deficit under the fiscal rule to 2029 (Reuters, 16 July 2026). In plain English: the fiscal rule is being deferred, not met. State debt is projected at 21.7% of GDP in 2027, from 19.9% in 2026, above the 20% level authorities regard as safe. Russia’s debt is low by the standard of other major economies, so this is not a solvency story. It is a signal about direction and about which levers are left. New tax rises are planned for 2027, including a windfall tax on metals and mining companies expected to raise about 200 billion roubles a year. Meanwhile the 2026 oil and gas revenue estimate was cut to 7.6 trillion roubles from 8.9 trillion (Reuters, 28 September 2026). Why did the central bank hold at 14%? Because the arithmetic of the budget is inflationary. A wider deficit financed partly by borrowing, with a bigger defence line, adds demand, and the central bank has said inflation is running above its path. Holding at 14% leans against that, but a 14% policy rate also raises the cost of the debt the Treasury is issuing, and the cost of capital for the civilian industry the index scores under the productive engine. The bank’s own October statement is explicit that a change in fiscal parameters may require a change in monetary policy (TASS, 23 October 2026). A 2027 baseline average rate of 12.5–13.5% means the cheaper money that would ease the squeeze is pushed out. Growth of 0.5–1% in 2026 is the cost. It is not collapse, and the figures here do not support that word. They support a slower, more specific claim: a war economy [continues on the page]

South China Sea China Philippines incidents October 2026

Source: https://worldwar3.army/reports/velocity-2026-W42.html

Philippine monitoring detected 101 Chinese vessels in September 2026, nearly double August's count. Six confrontations in ten days. Velocity, not presence.

Velocity Report 2026-W42 Grey-zone velocity: South China Sea, October 2026. Philippine monitoring detected 101 Chinese vessels near disputed features in September 2026, nearly double the 56 counted in August. Six dangerous engagements followed in the first ten days of October. This is not a spike. It is a shift from episodic pressure to sustained presence, and velocity — the rate at which the baseline changes — is what compresses strategic timelines and forces binary choices about treaty thresholds that were designed never to be tested. 12 OCT 2026 8 min Chinese vessels, Sep 2026 101 Monitored near Philippine-held features. Up from 56 in Aug (Philippine Star, 8 Oct 2026). Dangerous engagements, Oct 1–10 6+ Across Scarborough, Second Thomas, Thitu and Luzon Strait (USNI News, 9 Oct 2026). Increase in Chinese force events, Q2 2026 ~300% Apr–Jun vs Jan–Mar, per ACLED South China Sea Monitor. China official defence budget, 2026 +7% To 1.91 trillion yuan (~$277bn); Pentagon estimates real spending 40–90% higher (Defense News, 10 Mar 2026). What moved this week: grey-zone velocity, not events The number that matters is not 101. It is the fact that 101 is nearly double 56, and the doubling happened in a single month. Philippine Navy monitoring data released on 8 October showed that China deployed 101 coast guard, naval and maritime militia vessels near Thitu Island, Second Thomas Shoal and other Philippine-held features in September 2026, compared to 56 in August. That is not a presence. It is an acceleration. The incidents that followed were not unrelated. Between 1 and 10 October, Philippine and Chinese forces engaged dangerously at least six times across Scarborough Shoal, Second Thomas Shoal, Thitu Island and the Luzon Strait, according to USNI News tracking. On 4 October, a China Coast Guard vessel attempted to block Philippine fishers near Thitu. Key facts: sourced claims for citation Philippine monitoring detected 101 Chinese vessels near disputed features in September 2026, nearly double the 56 counted in August (Philippine Star, 8 October 2026). Philippine and Chinese forces engaged dangerously at least six times across Scarborough, Second Thomas and Thitu in the first ten days of October 2026 (USNI News, 9 October 2026). The US-Philippines Mutual Defense Treaty applies to armed attacks on Philippine coast guard vessels anywhere in the South China Sea, per 2023 Bilateral Defense Guidelines (US State Department, February 2026). ACLED recorded a 300% increase in events involving Chinese forces in the South China Sea between April and June 2026 compared to the prior three months. China's official 2026 defense budget rose 7% to 1.91 trillion yuan ($277 billion), though Pentagon estimates place real spending 40–90% higher at $388–526 billion (Defense News, 10 March 2026). The pattern is not new, but the velocity is. ACLED's South China Sea Monitor recorded a roughly 300% increase in events involving Chinese forces between April and June 2026 compared to the three months prior. What was episodic is becoming persistent. What was spectacular — encirclement exercises, large-formation transits — is being replaced by constant low-grade pressure that changes the baseline without offering a clear trigger for response. That is the entire grey-zone playbook, and it works precisely because treaty thresholds are binary while the approach to them is continuous. An armed attack triggers the US-Philippines Mutual Defense Treaty. Ramming, water cannons, blockades and the installation of uncrewed buoys do not, even when they achieve the same strategic effect: the erosion of access, the narrowing of options, and the repricing of what counts as normal. Why acceleration matters more than presence Static presence can be absorbed. Velocity cannot. A standing fleet of 50 vessels is a problem that can be monitored, worked around, and factored into routine operations. A fleet that doubles in a month is a compression of the decision timeline, because it signals that this month's baseline will not be next month's, and the defender has to decide now whether the trajectory is tolerable or whether waiting makes the response harder. This is window logic at the tactical level. A power whose relative position is deteriorating on a known schedule faces a choice between acting while it still can and accepting a permanently worse settlement later. The same logic applies to grey-zone creep. If Chinese presence is growing at this rate, and if that presence translates into de facto control — the ability to block, inspect, delay or turn back Philippine vessels without firing a shot — then the window for contesting that control is closing, and the cost of action rises each month it is deferred. The August-to-September data makes the velocity legible. It is no longer a question of whether China is increasing pressure in the South China Sea. It is a question of how fast, and whether the rate itself is accelerating. And the answer, as of early October 2026, is that it is. The treaty arithmetic: where persistence becomes provocation The US-Philippines Mutual Defense Treaty, signed in 1951, commits both nations to respond to an armed attack on either party's forces. Bilateral Defense Guidelines issued in 2023 and reaffirmed in a February 2026 strategic dialogue make the application explicit: the treaty covers armed attacks on public vessels, aircraft or armed forces — including coast guards — anywhere in the South China Sea. That clarity is both a deterrent and a constraint. It deters open attack, which is why China has not sunk a Philippine vessel. But it also creates a floor beneath which an enormous amount of coercion can occur without triggering the treaty. Ramming is not shooting. Water cannons are not missiles. Blocking access is not an armed attack, even when it achieves the same outcome: the denial of Philippine sovereign rights in its own exclusive economic zone. The grey area between legal harassment and armed attack is where the entire contest is being fought, and China's strategy is to occupy that space so thoroughly that by the time an incident crosses the threshold, the baseline will have shifted so far that a treaty response looks disproportionate. If 101 vessels is normal, and if ramming is normal, and if Philippine ships are routinely turned back without firing a shot, then what exactly is the armed attack that triggers Article IV? And if the threshold is never crossed, what is the treaty defending? This is not hypothetical. In the first ten days of October, Philippine and Chinese forces engaged dangerously at least six times, including a China Coast Guard attempt to block Philippine fishers near Thitu on 4 October. None of it invoked the treaty. The line held, but it was tested, and the fact that it was tested in the context of 101 vessels rather than 56 is what makes the velocity the story. China's closing window and the logic of now The acceleration is not random. It maps directly onto China's demographic and fiscal timeline . China's population fell by 3.39 million in 2025, the fourth consecutive annual decline, to 1.405 billion. Its fertility rate in 2025 was [continues on the page]

Demographic Decline and Closing Windows: Russia & China 2026

Source: https://worldwar3.army/reports/velocity-2026-W41.html

Russia recruited 232,000 troops while losing 240,000 in seven months. China's fertility hit 1.03. When demographic windows close, capability arithmetic changes.

Velocity Report 2026-W41 Demographic decline and closing windows: Russia and China, 2026. Russia recruited 232,000 contract soldiers between January and July 2026. It lost 240,000. China's fertility rate is 1.03 children per woman, and its population fell by 3.39 million in 2025. These are not policy problems. They are arithmetic problems with demographic clocks attached, and both leaderships can read the calendar. Published 5 October 2026 Key facts Russia recruited approximately 232,000 contract personnel between January and July 2026 while suffering about 240,000 casualties, including at least 140,000 irreversible losses ( ISW , July 2026). Russian births fell 6% in Q1 2026 to roughly 272,000, a two-hundred-year low (Russian official statistics, Q1 2026). Russia produced over 7 million artillery shells, mortar rounds and rockets in 2025, up seventeenfold from pre-invasion levels ( Estonian Foreign Intelligence Service , February 2026). China's population fell by 3.39 million in 2025, the fourth consecutive annual decline, to 1.405 billion (China National Bureau of Statistics, 2026). China's total fertility rate in 2026 is 1.03 children per woman, far below the replacement level of 2.1 ( Worldometers , 2026). PLA aircraft entering Taiwan's ADIZ fell 52% in the first seven months of 2026, with 1,200 sorties against 2,501 in the same period of 2025 ( Taiwan Ministry of National Defence , August 2026). Russian military losses surpassed 1.5 million killed and wounded, with approximately 300,000 casualties in the first nine months of 2026 (Ukrainian intelligence, October 2026). What moved this week, and which decline velocities it bears on Four numbers define week 2026-W41, and all four are velocity indicators rather than events. Russia's casualty-to-recruitment gap widened to net negative through July — 240,000 casualties against 232,000 recruited — while shell production hit 7 million rounds annually in 2025, seventeen times pre-invasion output. China's population fell 3.39 million in 2025, the fourth straight annual decline, while PLA aircraft sorties into Taiwan's air defense identification zone dropped 52 percent in the first seven months of 2026 compared with the same window in 2025. Russia's Q1 2026 birth rate hit 272,000, down 6 percent year-on-year and the lowest figure in two centuries. These are not discrete weekly events. They are the latest readings on trajectories that have been visible for quarters or years, and what makes them consequential this week is that they narrow the remaining timeline measurably. A power losing people faster than it can replace them is not in a stable equilibrium. A power with a fertility rate of 1.03 and a shrinking military-age cohort is managing a closing window whether it names it publicly or not. The claim here is not that either power will act inside that window. The claim is that the window itself — the span of years during which capability and demography still permit large-scale action — is closing faster than the casualty figures alone would indicate, and both leaderships have access to the same birth statistics the rest of us do. Russia's manpower arithmetic: burning faster than replenishing Russia's total military losses since the February 2022 invasion exceed 1.5 million killed and wounded, according to Ukrainian intelligence assessments released in October 2026. Roughly 300,000 of those casualties occurred in the first nine months of 2026 alone. In parallel, the Institute for the Study of War assessed in July that Russia had recruited approximately 232,000 contract personnel between January and July while suffering about 240,000 casualties in the same period, including at least 140,000 irreversible losses. The recruitment target for 2026 was approximately 409,000, but Moscow was not on track to meet it as of mid-year. That gap — 8,000 more losses than recruits in seven months — is not wide enough to force immediate mobilization, but it is wide enough to make the arithmetic visible. And the arithmetic gets worse every year the birth rate stays where it is. Russian births in Q1 2026 totaled roughly 272,000, down 6 percent from the prior year and the lowest quarterly figure in two hundred years. The replacement cohort for the losses Russia is taking today would have been born in the mid-2000s, when the birth rate was higher but still well below replacement. The cohort that would replace losses five years from now is being born right now, at 272,000 per quarter, and those numbers do not support a multi-year war of attrition at current casualty rates. There is a second number that matters here, and it cuts the other way. Russia produced over 7 million artillery shells, mortar rounds and rockets in 2025, according to the Estonian Foreign Intelligence Service — an increase of seventeen times compared with pre-invasion levels and up from 4.5 million in 2024. Production in 2025 included 3.4 million howitzer shells, 2.3 million mortar rounds, 0.8 million rounds for tanks and infantry fighting vehicles, and 0.5 million rockets for multiple-launch systems. Moscow also received between 5 and 7 million shells from Iran and North Korea since 2023, with North Korean ammunition accounting for up to half of all shells used in the second half of 2025. So Russia can scale industrial production. It cannot scale population. The question that matters for pricing the window is which constraint binds first: the ability to produce ammunition or the ability to field personnel who can use it. The current data points toward manpower as the binding constraint, which means Russia's strategic timeline is set by the demographic calendar rather than the industrial one — and that calendar is legible, published, and not subject to revision. China's demographic constraint: the military-age cohort that isn't coming China's population fell by 3.39 million in 2025, according to the National Bureau of Statistics, bringing the total to 1.405 billion. That was the fourth consecutive year of decline. The 2026 fertility rate is 1.03 children per woman, with a birth rate of 6.1 per 1,000 population — among the lowest in the world and far below the replacement level of 2.1. China's median age is 41.7 years, against a global average of 31, and the age structure shows 16.48 percent under 15, 69.4 percent aged 15–64, and 14.11 percent over 65. The demographic momentum is unambiguous: the military-age cohort is shrinking every year, and no policy reverses that inside a decade. Against that backdrop, PLA aircraft incursions into Taiwan's air defense identification zone fell 52 percent in the first seven months of 2026 compared with the same period in 2025. Taiwan's Ministry of National Defence recorded 1,200 PLA aircraft inside the zone between January and August 2026, down from 2,501 in the equivalent window of 2025. September 2026 saw 181 incursions, continuing the downward trend but still representing persistent activity. On September 5, Taiwan detected 20 sorties including J-11, Su-30, H-6, Y-8 and KJ-500 aircraft, with 12 crossing the median line into Taiwan's northern, central and southeastern ADIZ. Naval activity remained elevated: six PLAN vessels and five official ships were detected the same day. The reduction in air sorties has been read in some quarters as [continues on the page]

#	Page	URL	Words
1	Relative Decline Index: 20 Nations Scored \| worldwar3.army	https://worldwar3.army/	1588
2	The five engines of decline \| The Relative Decline Project	https://worldwar3.army/theory.html	1234
3	Method — every weight, every source \| The Relative Decline Project	https://worldwar3.army/method.html	881
4	World War 3 and Relative Decline: The Theory	https://worldwar3.army/conflict.html	1760
5	Who outlasts whom \| The Relative Decline Project	https://worldwar3.army/outlast.html	1557
6	United States decline: debt, interest and the dollar	https://worldwar3.army/united-states.html	1751
7	China: the decline nobody has priced \| Decline Project	https://worldwar3.army/china.html	1987
8	Russia — the second life, ending \| The Relative Decline Project	https://worldwar3.army/russia.html	2175
9	Europe: the continent that stopped compounding	https://worldwar3.army/europe.html	2323
10	Southeast Asia and Australia — the quiet winners \| The Relative Decline Project	https://worldwar3.army/asia-pacific.html	1453
11	The new frontiers — Greenland, the Arctic, the sea routes \| The Relative Decline Project	https://worldwar3.army/frontiers.html	1487
12	Africa — improving, and being bought \| The Relative Decline Project	https://worldwar3.army/africa.html	1278
13	South America — the middle-quality institutional trap \| The Relative Decline Project	https://worldwar3.army/south-america.html	1194
14	About, standards and corrections \| The Relative Decline Project	https://worldwar3.army/about.html	790
15	Russia Defence Spending 2027 and Budget Deficit	https://worldwar3.army/reports/velocity-2026-W44.html	1950
16	South China Sea China Philippines incidents October 2026	https://worldwar3.army/reports/velocity-2026-W42.html	2327
17	Demographic Decline and Closing Windows: Russia & China 2026	https://worldwar3.army/reports/velocity-2026-W41.html	3043

worldwar3.army: full text for AI agents