Candidate examples — historical collective action failures

Research list for the course, to be cut down. Contemporary cases (AI, climate, current politics) are covered separately and are not here.

Twenty-eight candidates, roughly ordered small scale to large. Module names refer to arc.md. Where a case is popular but historically contested, the entry says so and says what the dispute is — several of the best-known "collapse caused by collective action failure" stories are among the weakest evidentially, and using one without the caveat is a hostage to fortune with this audience.

Uncertainty is marked inline rather than smoothed over. Anything marked unverified needs a check before it goes on a slide.

Small groups

The Maxi Trial and the breaking of omertà

What happened: Tommaso Buscetta, a Sicilian Mafia member arrested in Brazil and extradited to Italy in 1984, began talking to magistrate Giovanni Falcone. His testimony, and that of other pentiti who followed, produced the Maxi Trial in Palermo (1986–87), which indicted several hundred defendants and convicted the majority; many convictions were confirmed on final appeal in 1992.
Why it's a collective action case: The literal prisoner's dilemma, including the payoff structure the name comes from. Omertà is a cooperative equilibrium held in place by the certainty of punishment for defectors — including punishment of relatives. Once the state could offer protection and reduced sentences, and once one person talked, the incentive to be the next to talk rather than the last rose sharply.
Could serve: The games; Some basic game theory; Repeated interaction and the folk theorem (what happens when the punishment mechanism is disabled).
Media: Substantial. The Maxi Trial was filmed in a purpose-built bunker courtroom and footage exists. Marco Bellocchio's Il traditore (2019) is a feature film about Buscetta. Photographs of the courtroom cages are widely reproduced.
Source: Alexander Stille, Excellent Cadavers (1995); John Dickie, Cosa Nostra: A History of the Sicilian Mafia (2004).

The Donner Party

What happened: A wagon party of around 87 emigrants was trapped by snow in the Sierra Nevada over the winter of 1846–47 after taking an untested shortcut. Roughly half died. The group fragmented into family units, and the surviving accounts include refusals to share food, charging for provisions, and a killing at Alder Creek, alongside real acts of sacrifice.
Why it's a collective action case: Small-group cooperation under acute scarcity, where the group was too large for a single family's loyalty to cover and lacked any authority able to enforce sharing. Useful precisely because it is mixed: cooperation held within families and failed between them, which is the kin-selection boundary made visible.
Could serve: Introduction to collective action; The cooperative ape; Olson (the family as the small privileged group inside a latent larger one).
Media: Ric Burns, The Donner Party (PBS American Experience, 1992). Survivor narratives and Patrick Breen's diary survive as primary sources.
Source: Ethan Rarick, Desperate Passage (2008). Note that details of cannibalism and of specific accusations vary between survivor accounts and have been argued over; the diary is the firmest source.

The trench truces of 1914–18 and their suppression

What happened: Along parts of the Western Front, opposing units developed tacit non-aggression — shelling at predictable times and places, not firing during rations, the Christmas 1914 fraternizations. High commands on both sides broke these up deliberately, using raids, unit rotation and orders requiring aggressive action.
Why it's a collective action case: A repeated prisoner's dilemma solving itself through the shadow of the future between units that faced each other for months, and then being destroyed by a third party with an interest in defection. It is the cleanest illustration that repetition and stable pairing are what does the work, since removing them removed the cooperation.
Could serve: Repeated interaction and the folk theorem; War; The evolution of cooperation (Axelrod uses this case himself).
Media: Joyeux Noël (2005) is a feature dramatization of the Christmas truce. Contemporary photographs of the 1914 fraternization exist and are in the Imperial War Museum collection. Letters and diaries are extensively published.
Source: Tony Ashworth, Trench Warfare 1914–1918: The Live and Let Live System (1980). Axelrod treats it as a case study in The Evolution of Cooperation (1984), ch. 4.

The Nottinghamshire split in the 1984–85 miners' strike

What happened: The UK National Union of Mineworkers struck from March 1984 against pit closures, without a national ballot. Most Nottinghamshire miners, whose pits were relatively profitable and less immediately threatened, kept working; the area later broke away to form the Union of Democratic Mineworkers in 1985. The strike ended in March 1985 with no agreement.
Why it's a collective action case: Olson's problem in its original industrial form. The benefit of a won strike is a public good across the union; the cost falls on the individual striker; and the sub-group whose pits were safest had the weakest incentive. The absence of a ballot removed the legitimacy that might have substituted for selective incentives.
Could serve: Olson; The games (free riding); Introduction to collective action.
Media: Extensive. BBC and ITV news archive, contemporary photojournalism (the Orgreave confrontation is heavily photographed), and documentaries including Still the Enemy Within (2014). Pride (2014) is a feature film on a solidarity campaign during the strike.
Source: The strike's historiography is politically contested. Andrew Richards, Miners on Strike (1996) is a standard academic account; Seumas Milne, The Enemy Within (1994) is sympathetic to the NUM leadership. The Nottinghamshire facts above are not in dispute; the interpretation is.

The Ik of Uganda — flagged as contested

What happened: Colin Turnbull's The Mountain People (1972) described the Ik, displaced from their hunting range by the creation of Kidepo Valley National Park and hit by famine, as having abandoned sharing, care of the old and care of children.
Why it's a collective action case: Turnbull's claim is that cooperation is a luxury that scarcity strips away — that the norms holding a society together are contingent on surplus. If true, it is a strong case. It is substantially disputed. Critics found that several of Turnbull's key informants were not Ik, that he imputed motives he could not have observed, and that his picture reflected a specific famine and displacement rather than a settled culture. Later fieldwork — Cathryn Townsend's work on generosity among the Ik — reports extensive sharing norms.
Could serve: Could serve The cooperative ape as a counter-case, but only with the dispute stated. Probably better used as an example of how attractive collapse narratives get, which is a live issue for this audience.
Media: Peter Brook staged The Ik (1975) from the book. A 2021 BBC radio documentary revisits the case. (Exact BBC title and date unverified.)
Source: Colin Turnbull, The Mountain People (1972); Cathryn Townsend et al., "Generosity among the Ik of Uganda" (Evolutionary Human Sciences, 2020); the open letter "Setting the record straight on The Mountain People" at humangenerosity.org.

Commons and resources

The Grand Banks cod collapse

What happened: Northern cod off Newfoundland had been fished for roughly 500 years. Catches rose sharply after the arrival of factory freezer trawlers in the 1950s–60s, peaking around 1968, then fell. Canada declared a moratorium on 2 July 1992. It put tens of thousands of Newfoundlanders out of work — the usual figure given is around 30,000–35,000 — and the stock has still not recovered to pre-collapse levels.
Why it's a collective action case: Subtractable commons with poor observability and a large, partly foreign fleet. Also a failure of the science–politics interface: quota advice was repeatedly set above what the stock assessments supported, and the assessments themselves were biased upward by using commercial catch rates from an increasingly efficient fleet.
Could serve: Ostrom and the commons (as the failure against which the successes are read); Did it ever work?; The games.
Media: Good. CBC has extensive archive including footage of Fisheries Minister John Crosbie confronting fishermen in Bay Bulls in 1992. Photographs of laid-up trawlers and empty plants are abundant.
Source: Ransom Myers et al. on the assessment failures (several papers, 1996–97); Alan Christopher Finlayson, Fishing for Truth (1994); Canadian DFO stock assessments.

The Peruvian anchoveta collapse

What happened: The Peruvian anchoveta fishery grew from almost nothing in the early 1950s to the largest single-species fishery in the world, peaking at roughly 12 million tonnes around 1970–71 (sources differ on which year). Catches fell by more than half in 1972 and again in 1973. A strong El Niño triggered the collapse; sustained fishing pressure on a stock already showing reduced biomass determined the depth of it.
Why it's a collective action case: Open access followed by heavy overcapitalization — a fleet and processing capacity sized for a catch the stock could not sustain. The distinctive feature is the interaction of a natural shock with an over-built fleet: the shock was not the failure, the inability to stop fishing during it was.
Could serve: Ostrom and the commons; Did it ever work?; Why this gets harder at scale.
Media: Photographs of the Chimbote fleet and fishmeal plants exist. No well-known documentary identified.
Source: FAO fishery statistics; M. H. Glantz's work on El Niño and the anchoveta; Wikipedia's "1972 Peruvian anchoveta crisis" collects the figures with citations. (A single authoritative monograph has not been checked; verify the peak-year figure before use.)

Soviet illegal whaling and the IWC

What happened: From 1948 the Soviet whaling fleet ran a systematic global campaign of illegal whaling, keeping two sets of books — falsified returns to the Bureau of International Whaling Statistics, true figures retained internally. Alexey Yablokov published the first account of the true scale in 1994. For 1948–79 the corrected North Pacific figures differ from those reported to the IWC by on the order of 20,000–26,000 whales, and figures were falsified for 8 of 12 hunted species. At least 17 blue whales were taken illegally in the first year of the international observer scheme.
Why it's a collective action case: A treaty regime with agreed quotas and no adequate monitoring. Everything the diagnostic asks about observability fails at once: defection was invisible, so it could not be punished, so the quota was not a constraint. The motive was internal — fleet production targets in the planned economy — which makes it a nested collective action failure: the Soviet planning system's own incentives cut against the international agreement.
Could serve: Did it ever work? (as the counterweight to the treaties that worked); Schelling (monitoring and verification); Coercion.
Media: Photographs of Soviet factory ships exist. Popular accounts in Pacific Standard and Hakai Magazine. No major documentary identified.
Source: Yulia Ivashchenko and Phillip Clapham, "Soviet Illegal Whaling: The Devil and the Details" (Marine Fisheries Review); Ivashchenko et al., "Soviet catches of whales in the North Pacific: revised totals" (Journal of Cetacean Research and Management).

The near-extinction of the American bison

What happened: Bison numbering in the tens of millions on the Great Plains were reduced to a few hundred wild animals by the mid-1880s, principally by commercial hide hunting in the 1870s after tanning processes made bison hide commercially valuable, and with railroads providing transport. US military and political actors welcomed the destruction as a way of subduing Plains nations.
Why it's a collective action case: Open-access slaughter with no property rights and no enforceable restraint, where each hunter's restraint would simply have left the animal for the next hunter. A Kansas bill to limit the killing passed the legislature and was reportedly not signed. (This last detail is repeated widely and should be verified before use.)
Could serve: The games (commons); Ostrom and the commons; The bill — the destruction was in part deliberate policy against the Plains nations, which makes it a case where the "failure" was someone's intended outcome.
Media: Strong. Photographs of bison bone piles (the Michigan Carbon Works image from around 1892) are among the most-used images in this whole area. Ken Burns, The American Buffalo (PBS, 2023).
Source: Andrew Isenberg, The Destruction of the Bison (2000); Dan Flores, American Serengeti (2016).

The passenger pigeon

What happened: Passenger pigeons were possibly the most numerous bird in North America, with flocks described in the billions. Commercial netting and shooting for market, aided by telegraph and rail, destroyed the population over roughly the second half of the nineteenth century. The last confirmed wild bird was shot around 1900; Martha, the last captive individual, died at Cincinnati Zoo on 1 September 1914.
Why it's a collective action case: Open access again, with a twist worth the slide: the species appears to have required very large colonies to breed successfully, so harvesting pushed it past a threshold rather than down a gradient. Individual restraint was not merely unrewarded, it was irrelevant below the threshold.
Could serve: The games; Introduction to collective action (a short, vivid case); Ostrom and the commons.
Media: Photographs of Martha and of netting operations exist. Extensive nineteenth-century press accounts.
Source: Joel Greenberg, A Feathered River Across the Sky (2014); A. W. Schorger, The Passenger Pigeon (1955).

The Dust Bowl

What happened: Plowing of southern Plains grassland accelerated through the 1910s–20s on high wheat prices. Drought from 1931 combined with exposed soil to produce severe dust storms across Oklahoma, Texas, Kansas, Colorado and New Mexico through the 1930s, with the worst — "Black Sunday" — on 14 April 1935. The Soil Conservation Service was created in 1935.
Why it's a collective action case: Soil erosion is not confined to the field that caused it, so conservation on one farm is a public good for the neighborhood. With prices falling, each farmer's rational move was to plow more, which made the aggregate worse. Also a case where the eventual fix was federal — mutual coercion, mutually agreed on.
Could serve: Introduction to collective action; Coercion; Changing the payoffs; The games.
Media: Exceptionally strong. Pare Lorentz, The Plow That Broke the Plains (1936); Ken Burns, The Dust Bowl (2012); the FSA photographic archive (Lange, Rothstein) is public domain.
Source: Donald Worster, Dust Bowl: The Southern Plains in the 1930s (1979); Timothy Egan, The Worst Hard Time (2006).

The Aral Sea

What happened: From the early 1960s Soviet planners diverted the Amu Darya and Syr Darya to irrigate cotton in Uzbekistan and Kazakhstan. The Aral Sea, once the fourth-largest lake in the world, lost most of its volume; by the 2000s it had split into separate bodies and the fishery was gone. The Kazakh North Aral was partly restored after the Kok-Aral dam (completed 2005).
Why it's a collective action case: Upstream–downstream river allocation, which became an international problem the moment the USSR dissolved in 1991 and a single planner's decision turned into a five-country bargaining problem with no enforcement. Note that the initial destruction was not a coordination failure at all — it was a deliberate, centrally chosen trade-off — which makes it useful for showing that central authority is not automatically the fix.
Could serve: Coercion (its limits); Polycentricity; Why the current stack can't hold.
Media: Strong. Satellite image sequences from NASA are public domain and widely used. Photographs of ships stranded in desert near Moynaq are iconic. Several documentaries exist.
Source: Philip Micklin's published work on the Aral (several papers, including in Annual Review of Earth and Planetary Sciences, 2007); NASA Earth Observatory.

Salinization in southern Mesopotamia

What happened: Thorkild Jacobsen and Robert McCormick Adams argued in 1958 that irrigation without adequate drainage progressively salinized soils in southern Iraq from roughly 2400 BC, forcing a shift from wheat to salt-tolerant barley, driving down yields, and contributing to the decline of southern centers relative to the north.
Why it's a collective action case: Drainage is a public good across an irrigation system — one farmer's drainage benefits neighbors, one farmer's over-watering damages them — and the damage accrues over generations while the benefit of over-irrigating is immediate. Fallowing to control salt costs the individual now and pays the group later.
Could serve: Why this gets harder at scale; The bill; Introduction to collective action.
Media: Cuneiform administrative texts recording yields are the primary source and photograph well. No documentary identified.
Source: Jacobsen and Adams, "Salt and Silt in Ancient Mesopotamian Agriculture," Science 128 (1958), 1251–58. The thesis has been challenged and defended since; treat the direction as well-supported and the strength of the causal claim as debated.

Deforestation of Iceland

What happened: Iceland was substantially wooded at the time of Norse settlement around 874 — the common figure is that birch woodland covered a quarter or more of the land area. Clearance for pasture and fuel, combined with sheep grazing that prevented regeneration and with volcanic soils vulnerable to wind erosion, reduced woodland cover to around 1% by the twentieth century.
Why it's a collective action case: Grazing on common land, with the twist that the damage is not to the grass but to tree regeneration — so nobody experiences the cost as their own loss in any given year. A slow commons tragedy, spread over centuries, on a bounded island with a literate legal culture that nonetheless did not regulate it.
Could serve: Ostrom and the commons; Why this gets harder at scale.
Media: Photographs of eroded highland and of reforestation plots. No major documentary identified.
Source: Andrew Dugmore et al. on Icelandic landscape change; the Icelandic Forest Service publishes the cover figures. (The pre-settlement cover figure varies by source; check before putting a number on a slide.)

Rapa Nui / Easter Island — flagged as contested

What happened: The standard story, popularized by Jared Diamond in Collapse (2005), is that Rapa Nui's inhabitants deforested their island competing to erect moai and collapsed into warfare and starvation before European contact.
Why it's a collective action case: On the standard telling, a status competition between lineages that consumed the shared resource base — an arms race in statues. It is the single most-cited example in popular collective action writing.
Contested, and the dispute matters. Terry Hunt and Carl Lipo argue the deforestation owed substantially to Polynesian rats eating palm seeds, that the population did not collapse before contact, and that the demographic catastrophe came with European disease and Peruvian slave raids in the 1860s. Recent modeling work (2025) supports a rat-plus-human account rather than a purely human one. Use it only with the dispute stated, or use it as a case study in how collapse narratives are selected.
Could serve: Introduction to collective action, but as a worked example of a contested case rather than a clean one; possibly The many names, and a critique of Moloch, since the Moloch reading of Rapa Nui is exactly the personification the course intends to argue with.
Media: Abundant photographs. Several documentaries, most of which tell the Diamond version.
Source: Jared Diamond, Collapse (2005); Terry Hunt and Carl Lipo, The Statues That Walked (2011) and "Ecological catastrophe, collapse, and the myth of 'ecocide' on Rapa Nui," in Questioning Collapse (2010).

Norse Greenland — flagged as contested

What happened: Norse settlements in Greenland, established around 985, were abandoned by the mid-fifteenth century. Diamond's account attributes this partly to failure to adapt — maintaining a European cattle-and-hay economy, not adopting Inuit marine technology, allocating scarce effort to the church.
Why it's a collective action case: On Diamond's account, elite incentives diverged from group survival: those who controlled the best farms had the least reason to change the system that made them elite. Contested. Later work emphasizes climate deterioration, sea-ice, the collapse of the walrus-ivory trade after African elephant ivory reached Europe, and evidence from isotope analysis that the Norse ate substantially more marine food than Diamond's account allows — that is, that they did adapt.
Could serve: The bill; Why the current stack can't hold — but only with the revisions, and possibly not worth the airtime given how much caveat it needs.
Media: Photographs of Hvalsey church and other ruins. Several documentaries, again mostly the Diamond version.
Source: Diamond, Collapse (2005); Jette Arneborg et al. on Norse diet isotopes; Kirsten Seaver, The Frozen Echo (1996).

Alliances, treaties and arms races

The Anglo-German naval race

What happened: Germany's naval laws from 1898 set out a battlefleet build-up; Britain responded, and HMS Dreadnought (1906) reset the competition by making existing battleships obsolete. British attempts at a negotiated building holiday — the Haldane Mission of 1912 among them — failed. Both states spent heavily and neither gained a relative advantage.
Why it's a collective action case: The textbook two-player arms race: mutual restraint dominates mutual building for both, and unilateral restraint is the worst outcome for whoever tries it, so both build. Both ended the race roughly where they started, poorer.
Could serve: War; The games; Schelling; The pattern beneath the metacrisis.
Media: Extensive photographs of dreadnoughts; naval launch newsreel footage from the period exists. Parliamentary debates and the naval laws are primary sources.
Source: Paul Kennedy, The Rise of the Anglo-German Antagonism (1980); Robert Massie, Dreadnought (1991).

The July Crisis and the mobilization schedules

What happened: Between the assassination of Franz Ferdinand on 28 June 1914 and the outbreak of general war in early August, each great power faced a mobilization decision in which the cost of mobilizing late was believed to be severe, because railway timetables could not be paused once begun. Russian partial then general mobilization, German ultimatum, and the invasion of Belgium followed within days.
Why it's a collective action case: A security dilemma with a first-mover advantage, made worse by a technology — the mobilization timetable — that removed the option of waiting. Everyone's decision was locally defensible and the aggregate was catastrophic. Note the contested historiography: A. J. P. Taylor's "war by timetable" is a strong version that later historians, notably Christopher Clark, have qualified heavily in favor of political choice.
Could serve: War; The pattern beneath the metacrisis; Schelling (commitment devices that work too well).
Media: Abundant photographs and newsreel. Diplomatic documents published in full.
Source: Christopher Clark, The Sleepwalkers (2012); A. J. P. Taylor, War by Timetable (1969) for the strong version.

Collapse of the Washington Naval Treaty system

What happened: The Washington Naval Treaty (1922) and London Naval Treaty (1930) capped capital ship tonnage by ratio between Britain, the US, Japan, France and Italy — a rare, working arms control agreement. Japan gave notice of withdrawal in 1934; the treaty system lapsed at the end of 1936 and naval building resumed without limit.
Why it's a collective action case: Worth using with the Anglo-German race because it shows cooperation actually achieved, with a specific mechanism — an observable, countable unit (capital ship tonnage) that made defection detectable — and then the specific reason it broke: an agreed ratio that a rising power would not accept as permanent. The fix failed on legitimacy, not on monitoring.
Could serve: Did it ever work?; Schelling; War.
Media: Photographs of the Washington conference and of ships scrapped under the treaty. Treaty texts are primary sources.
Source: Erik Goldstein and John Maurer (eds), The Washington Conference, 1921–22 (1994); the treaty texts.

The League of Nations over Manchuria and Abyssinia

What happened: Japan invaded Manchuria in September 1931; the League's Lytton Report (1932) found against Japan and Japan withdrew from the League in 1933, keeping Manchuria. Italy invaded Abyssinia in October 1935; the League imposed sanctions that excluded oil and did not close the Suez Canal, and Italy completed the conquest in 1936.
Why it's a collective action case: Collective security is a public good whose provision requires each member to bear a private cost — trade lost, war risked — to punish an aggressor who is not attacking them. Britain and France each had reasons to preserve relations with Italy against Germany. Nobody defended the norm, so the norm stopped existing.
Could serve: Did it ever work?; War; What modernity already did and Why the current stack can't hold (the enforcement gap in international institutions).
Media: Extensive newsreel, including Haile Selassie's address to the League Assembly in June 1936, which is filmed and is one of the strongest single clips available for this material.
Source: Zara Steiner, The Triumph of the Dark (2011); the Lytton Report.

Smoot–Hawley and the tariff spiral

What happened: The US Tariff Act of 1930 raised duties on a long list of imports. Canada, France and others retaliated; world trade contracted sharply over 1929–33, by around two thirds in nominal terms, though how much of that was tariffs rather than the depression itself is debated among economists.
Why it's a collective action case: Tariff setting is a prisoner's dilemma between states — mutual free trade beats mutual protection for both, unilateral opening is worst — and each round of retaliation is individually rational. Also useful for showing the solution, since GATT in 1947 is the institutional fix built directly on this experience.
Could serve: Changing the payoffs; Did it ever work?; The games.
Media: Photographs of Hoover signing; newspaper front pages. The economists' petition against the bill (1,028 signatories, 1930) is a good primary document.
Source: Douglas Irwin, Peddling Protectionism (2011). Irwin's own view is that the tariff's macroeconomic effect was modest, which is worth saying if the case is used.

Bank runs in the Great Depression

What happened: Waves of US bank failures ran from 1930 through to the national bank holiday of March 1933; several thousand banks failed. Federal deposit insurance was created by the Banking Act of 1933 and the runs largely stopped.
Why it's a collective action case: The purest coordination failure available, and structurally different from a prisoner's dilemma — the Diamond–Dybvig model has two equilibria, and which one you get depends only on what each depositor expects everyone else to do. Withdrawing is correct if others withdraw and wrong if they don't. Deposit insurance fixes it by making one equilibrium impossible rather than by changing anyone's character.
Could serve: The games (the dilemma/coordination/assurance distinction, which the arc flags as the one that does real work); Changing the payoffs; Coercion.
Media: Strong. Photographs of crowds outside banks; the run sequence in It's a Wonderful Life (1946) is a usable teaching clip and explains the mechanism in about ninety seconds.
Source: Douglas Diamond and Philip Dybvig, "Bank Runs, Deposit Insurance, and Liquidity" (Journal of Political Economy, 1983); Milton Friedman and Anna Schwartz, A Monetary History of the United States (1963).

Civic and civilizational

Greek disunity and Philip II

What happened: The Greek city-states, having exhausted each other in the Peloponnesian War (431–404 BC) and its successors, failed to combine against Macedon. Demosthenes' Philippics from 351 BC onward argued for exactly that and did not carry the point in time; Philip defeated the Athenian–Theban coalition at Chaeronea in 338 BC and imposed the League of Corinth.
Why it's a collective action case: Balancing against a rising power is a public good — everyone benefits if Philip is stopped, each city would rather someone else's hoplites did it — and Philip could pick them off in sequence and offer separate terms. Also a good illustration of why a large player changes the structure: no Greek city was big enough to act alone, which is question 5 of the diagnostic.
Could serve: War; Olson; Introduction to collective action.
Media: The Philippics survive in full and are quotable. Archaeological site at Chaeronea, including the Lion monument. No specific film identified.
Source: Demosthenes, Philippics and Olynthiacs; N. G. L. Hammond, Philip of Macedon (1994).

The liberum veto and the partitions of Poland

What happened: In the Polish–Lithuanian Commonwealth's Sejm, any single deputy could void a session's legislation. First exercised by Władysław Siciński in 1652 — though that first use was narrow, protesting a session extension, and did not undo the session's completed work. In the eighteenth century the practice became routine and foreign powers bribed deputies to use it. The Commonwealth was partitioned out of existence between 1772 and 1795.
Why it's a collective action case: A unanimity rule is a decision technology, chosen to protect against tyranny of the majority, which has the property that the cost of blocking falls on everyone but the blocker. It is the clearest historical case of a governance rule that converts a collective action problem into a permanent one, and it generalizes straight to the Security Council veto and to consensus requirements in modern treaty bodies.
Could serve: Polycentricity (what institutional design gets wrong); Coercion; Why the current stack can't hold.
Media: Paintings — Matejko's Rejtan (1866) depicts the 1773 partition Sejm. No documentary identified.
Source: Norman Davies, God's Playground: A History of Poland (1981); Jacek Jędruch, Constitutions, Elections and Legislatures of Poland (1982).

Norm collapse in the late Roman Republic

What happened: Sulla marched his legions on Rome in 88 BC — the first Roman commander to do so — after his command against Mithridates was transferred by a tribune's law. The precedent held. Within two generations the same move had been made repeatedly, and the Republic's institutions did not survive it.
Why it's a collective action case: Unwritten constitutional restraints are a public good sustained by everyone's expectation that everyone else will observe them. Once one player defects and profits, restraint by anyone else becomes unilateral disarmament — the return of a norm-following politician being exile or death. Nobody rationally goes second on restoring the norm.
Could serve: Repeated interaction and the folk theorem (what the equilibrium rests on); War; Coercion.
Media: Ancient sources — Plutarch's Sulla, Appian's Civil Wars — are usable primary text. Rome (HBO, 2005–07) is dramatization, not evidence.
Source: Plutarch, Life of Sulla; Appian, The Civil Wars; Mary Beard, SPQR (2015).

The Great Leap Forward and the communes

What happened: From 1958 Chinese agriculture was reorganized into large communes with collective work-point remuneration, alongside grain requisition based on inflated reported output and diversion of labor to backyard steel. Famine from 1959 to 1961 killed tens of millions; estimates range widely, commonly 15–45 million, and the range is itself the subject of active dispute.
Why it's a collective action case: Two failures stacked. The commune's remuneration system severed effort from reward at a group size far past where monitoring worked — Olson's problem imposed by policy. On top of it, a reporting failure: local cadres had strong private incentives to overstate harvests and none to report shortfall, so the center requisitioned grain that did not exist.
Could serve: Olson (group size and monitoring); Coercion (its limits, again); The bill.
Media: Chinese state propaganda footage and posters from the period are abundant and are themselves evidence of the reporting failure. Yang Jisheng's Tombstone draws on archives.
Source: Yang Jisheng, Tombstone (2008/2012); Frank Dikötter, Mao's Great Famine (2010). Both are contested on numbers; Dikötter's higher estimates in particular are disputed.

The Great Stink and London's sewers

What happened: London's sewage discharged into the Thames; cholera epidemics recurred through the 1830s–50s. Proposals for a unified sewer system were made repeatedly and not funded. In the hot summer of 1858 the smell reached the Palace of Westminster, and Parliament passed enabling legislation within about three weeks. Bazalgette's system was built from 1859.
Why it's a collective action case: Sanitation is the standard public good — non-excludable, expensive, and with the benefit diffused across a population and across future decades. It went unprovided for twenty years while everyone agreed it was needed, and was provided in weeks once the cost fell on the decision-makers personally. Useful as a solved failure, which the course arc specifically wants alongside the unsolved ones.
Could serve: Introduction to collective action (a failure that got fixed); Coercion; Changing the payoffs.
Media: Strong. Punch cartoons from 1858 ("The Silent Highwayman," "Father Thames Introducing His Offspring") are public domain and excellent slides. Photographs of the Crossness and Abbey Mills pumping stations.
Source: Stephen Halliday, The Great Stink of London (1999).

The 1953 North Sea flood

What happened: On the night of 31 January–1 February 1953 a storm surge broke sea defenses in the Netherlands, England and Scotland, killing roughly 1,800 in the Netherlands and over 300 in England, with further deaths at sea. Dutch dikes had been under-maintained through the occupation and post-war reconstruction, and known weaknesses had gone unfunded. The Delta Works followed.
Why it's a collective action case: Flood defense is a public good with deferred, invisible benefits and immediate, visible costs — maintenance always loses budget arguments until it doesn't. The Dutch water boards are one of Ostrom's better examples of long-lived commons institutions, which makes the failure more interesting, not less: a functioning institution was starved of resources by a higher level of government under competing priorities. That is a polycentricity failure specifically.
Could serve: Polycentricity; Ostrom and the commons; Coercion.
Media: Photographs and newsreel exist; Dutch commemoration archives are extensive. The Watersnoodmuseum holds primary material.
Source: Standard Dutch historical accounts of the Watersnoodramp; Ostrom's treatment of the Dutch water boards is in Governing the Commons (1990). (The specific claim about pre-1953 maintenance underfunding is well-established in Dutch sources but the English-language citation has not been pinned down — check before use.)

Notes on what is missing

Two structural gaps worth filling in a second pass:

Non-European small-scale cases. The small-group section leans Western. Pastoralist range disputes and irrigation-system failures outside Europe would balance it, and Ostrom's own case files contain failures as well as successes.
Failures of positive commons. Almost everything here is depletion of a subtractable resource. Undersupply cases — a research programme nobody funds, a standard nobody adopts — are the other half of the arc's distinction and are barely represented. Format wars and failed standards bodies are the obvious source, if a pre-2000 example can be found that doesn't read as contemporary tech.