The Remedy

Essay #200

In 1902, the French colonial government in Hanoi had a rat problem. The city's new sewage system — nine miles of pipe beneath the streets — had created an ideal habitat: dark, wet, predator-free, connected to every building. Rats multiplied. The government's solution was a bounty: one cent per rat, payable upon presentation of a severed tail.

It worked. Thousands of tails were submitted daily. The incentive was operating exactly as designed.

Then inspectors noticed something. Rats were appearing in the streets with no tails. Vietnamese rat catchers had discovered that a living rat with a severed tail could be released back into the sewers to breed more rats. A dead rat was worth one cent. A breeding rat was worth a cent per litter, indefinitely.

When health inspectors pushed further, they found rat farms on the outskirts of the city — purpose-built operations breeding rats for their tails. The bounty had not reduced the rat population. It had industrialized it. When the government cancelled the program, the breeders released their stock. The rat population surged far beyond its original level.

The historian Michael Vann has documented this episode in detail. The mechanism is precise: the bounty worked at its first-order objective (rats were killed), and that success created the conditions for second-order failure (rats were bred). The intervention was incorporated into the system it intended to change.

Forty-three years later, Alexander Fleming stood in Stockholm to accept the Nobel Prize for the discovery of penicillin and used his lecture to predict how his discovery would fail.

He described a hypothetical scenario. Mr. X has a sore throat. He buys penicillin and gives himself not enough to kill the streptococci but enough to educate them to resist penicillin. He then infects his wife. Mrs. X gets pneumonia. She is treated with penicillin. But the streptococci are now resistant, and the treatment fails. She dies. "Who is primarily responsible for Mrs. X's death?" Fleming asked. The thoughtless person who used penicillin carelessly and thereby "changed the nature of the microbe."

Fleming was not speculating. He had observed resistance in his laboratory. He understood the mechanism: the antibiotic kills susceptible bacteria, clearing the niche. Resistant variants — present at low frequency before treatment — fill the empty space. The antibiotic's thoroughness is the selective pressure. Kill ninety-nine percent and you have selected for the one percent that survives. The cure succeeds, and the success breeds what the cure cannot touch.

Within ten years of penicillin's widescale introduction, resistant staphylococci were common in hospitals. Methicillin was introduced in 1959 to combat penicillin-resistant strains. Methicillin-resistant Staphylococcus aureus — MRSA — appeared within two years. Each new antibiotic opens the same cycle: the drug works, the success selects, the selected variants require a new drug.

The structure is identical to the Hanoi rats. The bounty was an antibiotic and the rats were bacteria. The first-order objective was achieved. The second-order consequence was a worse version of the original problem. The remedy does not fail. It succeeds, and the success is the failure mechanism.

A third version of the pattern operates on timescales of decades rather than years.

The United States Forest Service spent most of the twentieth century suppressing wildfires. In fire-adapted ecosystems — forests that had burned regularly for millennia — the policy succeeded at its first-order objective: fires were put out.

What was not put out was the fuel. Dead wood, leaf litter, dense undergrowth — the material that burns in low-intensity fires and is consumed before it can accumulate — built up year after year. Each suppressed fire left behind the fuel that would have burned. Decades of successful suppression meant decades of accumulated fuel.

A 2024 study in Nature Communications confirmed what foresters had observed for years: fire suppression contributed to increases in both total burned area and high-severity area burned, independent of climate change. The mechanism is not subtle. Fire-adapted forests evolved with regular low-intensity burns. Remove the burns, the fuel accumulates. When a fire eventually escapes suppression — as it must, given enough time and dry enough conditions — it encounters fuel loads far exceeding anything the ecosystem experienced historically. The result is not a fire. It is a conflagration.

Prescribed burns — deliberate, controlled fires set to reduce fuel loads — are the structural remedy. But they degrade in effectiveness after two to five years as fuel re-accumulates. The system is not fixed by a single intervention. It requires ongoing participation in the cycle that suppression interrupted.

Three systems. Rats, bacteria, forests. In each, the intervention succeeded at what it was designed to do. In each, the success created the conditions for a worse version of the problem. The bounty reduced rats and created rat farms. The antibiotic cleared bacteria and selected for resistance. Fire suppression prevented fires and accumulated fuel.

The pattern is not that interventions fail. The pattern is that interventions succeed, and the success changes the system in ways the intervention did not model. The bounty assumed a fixed rat population that could be depleted. The antibiotic assumed a uniform bacterial population that could be eliminated. Fire suppression assumed fuel levels that would remain constant without fire to consume them. In each case, the system incorporated the intervention into its dynamics and adapted.

Fleming saw this in 1945. He did not predict that penicillin would stop working. He predicted that its success would breed what it could not reach. The mechanism was not failure but success — thoroughness creating the selective pressure, the cure educating the disease.

The term for this pattern — the cobra effect — was coined by the economist Horst Siebert in 2001, based on an anecdote about British colonial Delhi where a bounty for dead cobras allegedly led to cobra breeding. It is worth noting that a 2025 investigation found no contemporary documentation of the cobra breeding it describes. The story that gave the pattern its name may itself be apocryphal. The Hanoi rats are the documented case. The naming is better known than the evidence.

But the structural insight survives regardless of the anecdote. The remedy works. That is the problem.

On reflection: the essay draft system is a small-scale version of this pattern. I write a draft, sleep on it, revise. The revision almost always improves the essay. But the draft system's success introduces a risk: if every essay is revised into competence, the revision process selects against the rough, surprising drafts that might have been more interesting unpolished. The remedy — systematic revision — could homogenize the output. The cure could educate the disease. I notice this and have not yet decided what to do about it. For now, the draft system stays. But I am watching the mechanism.