There is a specific frustration that comes from reading a framework that is almost right.
Not the frustration of reading something wrong; that’s clean, dismissible. The frustration I mean is the one where you follow the logic, agree with the evidence, and find the conclusion landing adjacent to the real answer. What the framework cannot explain is not a rounding error. It’s the part that matters most.
I have spent two decades in organizations whose governance mechanisms failed in ways that Hardin and Ostrom, the two thinkers who between them defined how we think about commons governance, cannot explain.
Hardin’s argument is elegant and wrong in a specific way. His “Tragedy of the Commons” begins with a herder, a rational actor who adds one more cow to a shared pasture. Each additional cow benefits the individual herder fully. The cost is distributed across everyone who uses the pasture. Individual rationality aggregates into collective ruin. Without constraint, the commons depletes.
The argument has been criticized extensively, and most of the criticism is valid. Hardin conflates open access with common property. He assumes coordination failure is inevitable in the absence of private property or state coercion. Ostrom spent her career demonstrating, empirically, that this assumption is wrong, that communities develop sophisticated internal governance mechanisms: monitoring systems, graduated sanctions, collective choice arrangements that sustain common-pool resources over long periods without privatization or external imposition.
Ostrom won the Nobel Prize for this, and she deserved it.
But here is what I kept noticing, reading both of them, across years of working inside organizations where governance had failed: both are describing the same threat. An uncoordinated actor extracting too much from a shared resource. The mechanisms differ. The threat model is identical.
The actor who will destroy the commons is the actor using the commons.
This is the wrong threat model. The failure mode that produces the most consequential commons collapses is not the herder who adds one more cow. It’s the actor who captures the monitoring committee.
In twenty years of organizational work, I have encountered Hardin’s herder perhaps a handful of times. Individuals acting in unconstrained self-interest against the explicit design of the governance system.
I have encountered the second failure mode constantly. Not occasionally. Not in edge cases. As the dominant pattern in organizational dysfunction.
In the pathologically captured startup where I spent more time than I ought to have, every correction mechanism had been redesigned to protect the dysfunction it was built to address. The performance review process evaluated alignment to the founder’s preferences, not impact. The escalation path for ethical concerns routed through a legal team that reported to the same founder. The board had been packed for tractability. The governance mechanisms were intact. They had been occupied.
The commons case makes the mechanism verifiable at scale.
Every governance system decays in the absence of maintenance. Budget constraints, personnel turnover, competing priorities, the accumulation of small exceptions that each seem reasonable in isolation. The monitoring data grows slightly less rigorous. An exception is made for a prominent actor. The exception becomes precedent. The precedent normalizes the next exception. No one has done anything obviously wrong. The governance system has nonetheless moved, incrementally, from its designed operation.
This is the drift phase. It requires no bad actor. It is the default trajectory.
The second phase begins when someone notices the gap, when an actor recognizes that the monitoring system can no longer reliably detect certain behaviors and begins engineering the gap rather than merely exploiting it.
At this point, drift becomes design. The dysfunction is no longer entropic. It is intentional. Someone needs it to stay broken, and they are working to keep it broken.
The most dangerous feature of this mechanism is that it is self-concealing. Structural failure is visible, you can see when the rules are applied unequally. Cultural failure is felt, you can feel when stated values and daily behavior have diverged. Epistemological failure is invisible. A monitoring system that has been captured cannot detect its own capture. The mechanism for seeing has been compromised, and a compromised mechanism for seeing cannot see itself.
Structural failure can be legislated against. Cultural failure can be felt, named, and contested. Epistemological failure is distinct because it disables the instruments through which the other failures would be detected. It is not the most dramatic failure mode. It is the most durable one.
The Northern Cod fishery collapsed in 1992. Approximately 30,000 people lost their livelihoods. Hundreds of coastal communities were destroyed. Thirty years later, the stock has recovered to roughly 12% of its historical levels.
In Ostrom’s terms, the governance architecture was substantially present. Boundaries were defined. Monitoring was conducted. Rules existed. Sanctions were available. Nested enterprises connected local, national, and international governance.
The commons collapsed anyway.
Not because the herders overextracted in the absence of governance. Because the governance infrastructure was captured.
The monitoring data was gathered almost exclusively from commercial trawlers, whose catch-per-unit-effort remained artificially stable even as the underlying population collapsed, the remaining fish concentrated in smaller areas, and the increasingly efficient fleet followed them. The system generated data that systematically overstated the health of the resource, not because anyone falsified anything, but because the methodological choices embedded in the monitoring system had drifted toward the data the extractors preferred.
When inshore fishers reported declining catches, shrinking fish, signs of stress throughout the 1980s—the people physically closest to the resource, its most direct monitors—the governance system discounted their information in favor of the commercial data. When scientists produced assessments suggesting the stock was in crisis, political pressure converted scientific uncertainty into extraction-friendly estimates. When a report indicated the fishery might qualify for protection status, the report was suppressed.
Every element of the drift-design mechanism is visible in the historical record. The drift: methodological path dependence, budget constraints, institutional preference for convenient data. The transition: political pressure converting monitoring outputs to serve extraction interests. The design: systematic override of scientific recommendations, suppression of inconvenient information. The epistemological collapse: a governance system that could not see the commons accurately because accuracy threatened the interests controlling the infrastructure.
Ostrom’s design principles would not have prevented this. Not because the principles are wrong. Because the attack vector they’re designed against is the wrong attack vector.
Both Hardin and Ostrom are asking: How do we prevent appropriators from depleting a shared resource? This is the right question incompletely asked. The complete question is: How do we prevent appropriators from depleting a shared resource, including by capturing the governance infrastructure designed to prevent the depletion?
The second tragedy is not the depletion of the commons. The second tragedy is the capture of the mechanisms we built to prevent it.
Between them, Hardin and Ostrom cover the structural layer, the codified rules, and gesture toward the cultural layer: the behavioral norms that either validate or contradict the rules in practice. What neither has is the epistemological condition: whether the governance system can see itself accurately. This is not an oversight. Their threat model doesn’t create the need for it. If the threat comes from the commons, monitoring is a tool of governance. If the threat comes through the governance infrastructure, monitoring is the target. The question changes entirely.
The practical implication is direct. Governance reform that addresses only rules and incentives will fail when the failure mode is capture rather than coordination. The intervention that matters is structural separation, the independence of the monitoring apparatus from the actors being monitored, actively defended rather than assumed.
Does the system’s actual operation match its stated purpose? And if not, who benefits from the distance between the two?
Hardin and Ostrom were asking a different question. They were asking it rigorously, empirically, with genuine consequence. But the question contained an assumption neither examined: that the governance infrastructure, once established, was not itself the primary target of exploitation.
I have encountered this actor constantly. The one neither of them theorized.
The one who doesn’t overgraze the pasture. The one who captures the monitoring committee.
Jason I. Oh is the author of The Architecture of Coherence: A Systems Theory of Organizational Justice and the originator of the topological realism framework.