The previous essay ended on the limit of five historical cases. They showed what happens when the moral fragility inside a business ends up on its balance sheet: franchise death, settlement payments, a growth freeze, a business model destroyed, expropriation. But we chose those five cases because the damage had actually happened. Our failure sample can tell you how bad it gets.
Most writing in stops there and lets the reader assume the failures are representative. We think that assumption is the worst move in risk analysis, so we did a little more work. For each of the historical cases, we built a full cohort: every comparable entity carrying comparable exposure at the same starting date, selected by criteria we fixed in advance, before looking at outcomes. Then we followed everyone through time, including the firms nothing happened to, and coded each one at its worst outcome. Nothing (X0). Minor penalties (X1). A hit material to the entity, a forced exit, a multi-year constraint (X2). Bankruptcy, franchise death, expropriation (X3).
Five cohorts. Forty-one entities. To find out how often. Here is what the denominators say.
Nine for nine
Start with opioids. In May 2007, the month Purdue’s operating company plead guilty to federal misbranding charges, nine companies had material revenue from manufacturing or distributing prescription opioids: Purdue, Endo, Mallinckrodt, Johnson & Johnson, Teva, Allergan, and the three big distributors, McKesson, Cardinal Health, and AmerisourceBergen.
Follow all nine for eighteen years. Every one of them paid at the material level or worse. Nine for nine.
Three died or nearly did. Purdue: bankruptcy and dissolution. Endo: Chapter 11 plus a $1.086 billion criminal fine. Mallinckrodt: two bankruptcies and a $1.7 billion settlement obligation. The other six wrote large checks: about $5 billion and a forced exit from the US opioid business for Johnson & Johnson, $4.25 billion for Teva, $2.37 billion for Allergan, and $6.86 billion, $5.56 billion, and $5.58 billion for the three distributors, paid out over roughly eighteen years.
One more company entered this same trade but only after after our starting date, so it does not count in the nine. But the consequences were still there. Insys Therapeutics launched a fentanyl spray in 2012. Seven years later it was bankrupt, dissolved, and its founder was in federal prison.
For this cohort, in this window of time, a material happened to all. The only thing that varied was severity. The companies that were nothing but opioids went bankrupt. The diversified majors and the distributors wrote checks and survived.
Twenty percent
Now the contrast cohort, and the most important finding in this study.
We assembled ten cases from industries with documented public costs: the tobacco majors, three firearms makers (Remington, Smith & Wesson, Sturm Ruger), Coca-Cola and PepsiCo, Anheuser-Busch and Diageo, Caesars and MGM. Same fragility family as the opioid cohort. Comparable exposure on paper. Followed for thirty-one years.
Two of the ten took a material hit.
This wasn’t moral luck. It was just legal statute. In 2005, Congress passed the Protection of Lawful Commerce in Arms Act, which blocked states from running the tobacco playbook against gunmakers. In 2004, the House passed the “Cheeseburger Bill”, and state versions followed, shielding food companies from obesity recovery suits. Nobody ever found a good way to sue alcohol or gambling industries. The obesity suits against Coca-Cola and PepsiCo were dismissed. Caesars’ $18 billion bankruptcy was a leverage story, not a recovery of the harm.
The exception. Remington still died: two bankruptcies and a $73 million Sandy Hook settlement. A blocked channel to capture the full weight of the moral harm is just a political fact. A statute that blocks recovery can be repealed. The exposure is still there seeping.
Put the two cohorts side by side and what separated them was whether a working legal tool existed. Opioids faced one, inherited from tobacco. Firearms and food had Congressional protection. Whoever prices this class of risk on exposure alone will be wrong.
The certain layer
Seven US retail banks carried documented aggressive-sales or consumer-abuse cultures by the mid-2010s: Wells Fargo, TD Bank’s US operation, Santander Consumer, Bank of America, Regions, U.S. Bank, Fifth Third. Over twelve years, every single one of them paid regulatory penalties for it. $250 million, $191 million, $37.5 million, $20 million. Three of the seven took structural hits: Wells Fargo’s seven-year asset cap, TD’s $3.09 billion in penalties, guilty plea, and its own asset cap, and Santander’s roughly $550 million 34-state settlement.
The same appears in adult-content platforms after the card networks cut off Pornhub in December 2020. Six major platforms; five years; every one of them ended under penalty or restriction. Texas settlements, EU designations and minor-protection proceedings, state-by-state access blocks. The Supreme Court solidified it upholding Texas’s age-verification law in June 2025.
An event that happens to everyone is not a risk scenario. It is an operating cost. Companies with this cultural profile should carry the minor-penalty layer as an expected expense, the way a trucking company carries fuel. Treating it as a probability-weighted scenario understates a cost that reliably arrives. Save the scenario weights for the material tier, which hit roughly a third to two-fifths of the bank cohort in twelve years.
Who takes the worst outcome is not random. Six platforms carried near-identical category exposure. One of them carried a documented record of non-consensual content failures its competitors did not. That one took the terminal outcome; the other five took penalties and constraints. Same pattern as opioids, where the pure plays died and the diversified survived.
The flag and the precedent
Next: nine jurisdictions that held large stocks of foreign capital around 2010 under governments whose stability rests on coercion. Russia, Venezuela, Myanmar, China, Belarus, Kazakhstan, Saudi Arabia, Turkey, Vietnam.
Fifteen years later, six of the nine had forced material losses on foreign capital. Russia’s seizure decrees and more than $167 billion in foreign losses. Venezuela’s uncompensated expropriations and an $8.7 billion arbitration award. Myanmar’s post-coup forced exits, with Telenor’s unit sold for $105 million under junta steering. China’s tutoring ban, which stranded foreign investors in a sector worth around $100 billion. Belarus’s disposal freeze on 190 companies. Kazakhstan’s claims of up to $150 billion against the Kashagan developers.
We checked our own index against the cohort. All nine jurisdictions, the ones that hit foreign capital and the ones that did not, carry the coerced flag in the current Moral Disorder Index run. The coercion flag caught every jurisdiction that seized or stranded foreign capital.
Turkey and Vietnam carry the same flag, and nothing happened to foreign capital there in this timeframe. Vietnam, flagged, grew its realized FDI stock past $300 billion, and the US government’s own investment-climate reporting found no expropriations of foreign investment there.
What does separate the worst outcomes is precedent. Of the nine, exactly two had expropriated foreign or major domestic capital before our 2010 starting date: Russia (Yukos, 2003) and Venezuela (the 2007 wave). Both went terminal.
Bands, not points
Here is what we will not do with these numbers.
The cohorts are small: six to ten members each. The statistical intervals around a frequency like 9 of 9 or 2 of 10 are wide. Each cohort is also one draw of history: the nine opioid outcomes were correlated with each other, produced by one litigation wave, not nine independent coin flips. And the cohorts were assembled in 2026 by analysts who know how the stories ended. We fixed membership criteria before coding outcomes, which is the available discipline, but retrospective work cannot be fully blinded.
We have a forward fix of applying the same coding to entities whose outcomes are not yet known.
These are measured probabilities moving an assessment’s starting point. They do not decide its conclusion, and they are stated as ranges in everything we produce. What they replace is worse: gut feeling, or the silent assumption that the famous failures are the whole story.
Knowing which world you are in is worth a great deal of money, and the only way to know is to follow everyone.
This essay summarizes the matched-cohort study behind UnseenFront’s Moral NPV methodology, companion to “Priced at Zero.” Every outcome record is drawn from primary or high-quality secondary sources linked in the text. The MDI dataset and methodology (license: CC BY-NC-SA 4.0) are published on the Open Science Framework.