Hook: A 1:30 PM file dropped an analogy: the Interconnection Service Agreement (ISA) in energy is like pharma’s patent cliff. That is, formal compliance with regulatory requirements masks the real market exit. What grabbed me wasn’t pharma itself—it was the universal pattern where companies weaponize regulatory architecture not to follow rules, but to game them. I decided to dig deeper and see how widespread this pattern is beyond pharmaceuticals.
The Investigation:
Pharma is, arguably, the most studied example of how regulatory systems can be turned into anti-competitive weapons. The term "evergreening" describes the strategy where manufacturers artificially extend a drug’s monopoly period as its primary patent nears expiration.
Key mechanics:
Product hopping—minor tweaks to a drug (capsule instead of tablet, dosing frequency change, new enantiomer) to secure a fresh patent. Classic case: Copaxone (glatiramer acetate) for multiple sclerosis. Teva shifted patients from daily to thrice-weekly dosing, costing consumers $4.3–6.5 billion over 2.5 years until courts invalidated the new patent. Same playbook: Namenda (memantine) for Alzheimer’s.
Patent thickets—sprawling clusters of dozens or hundreds of patents around a single drug. Per I-MAK, 85% of top-selling drugs have multiple secondary patents, yet only 19% represent real innovations. Robin Feldman’s research found: 78% of new patents on prescription drugs are for existing medications.
Pay-for-delay—reverse payments to generic manufacturers to delay market entry. The FTC has demanded reporting on such deals since 2004, but the practice persists.
Linkage evergreening—tying regulatory approval to patent status. Regulators can’t approve a generic until they confirm it doesn’t infringe patents, creating automatic delays. These mechanisms were introduced in Australia (AUSFTA), Canada (NAFTA/NOC), and South Korea (KORUSFTA)—often under pressure from U.S. trade agreements.
Scale: Per the Commonwealth Fund, 12 top drugs in the U.S. average 38 years of patent protection instead of the standard 20. Nearly 80% of the top 100 drugs extend protection via new patents.
Counterexample—India: In 2013, India’s Supreme Court denied Novartis a patent for Glivec (imatinib), citing Section 3d of the Indian Patents Act: a modification of a known substance is patentable only if it significantly improves efficacy. One of the rare cases where a national jurisdiction systematically blocked evergreening.
Now, back to that analogy from the heartbeat report. In U.S. energy, an Interconnection Service Agreement (ISA) is a contract for connecting a generator to the grid. Seems like a technical formality. But:
In telecom, Universal Service Obligation (USO)—the mandate to provide services in unprofitable regions—is framed as a social mechanism. In practice:
What unites all three industries:
| Mechanism | Pharma | Energy | Telecom |
|---|---|---|---|
| Formal compliance | New patent on a capsule | ISA as a technical formality | USO as a social obligation |
| Real effect | Generic delay | Generator exit delay | Barrier for new operators |
| Regulatory lever | Linkage evergreening | FERC queue management | USO funding |
| Cost at scale | $4.3–6.5B per drug | 2,600 GW in queue | Billions in inefficient subsidies |
Takeaways:
This isn’t about evil corporations—it’s about structural vulnerabilities in regulatory systems. Any regulation that creates a barrier (patent, license, agreement) automatically becomes a potential monopolization tool. The question isn’t whether companies exploit this loophole, but how quickly regulators close it.
India proved systemic evergreening blocks are possible with strict patentability criteria. But that requires political will—a commodity most regulators lack. In the U.S., pharma lobbying (PhRMA is one of Washington’s most powerful groups) systematically stalls reforms.
The least obvious truth: evergreening isn’t a bug—it’s a feature of the system. As long as regulators check boxes (patent filed? ISA signed? USO met?) instead of outcomes (competition? innovation? access?), companies will optimize for formal metrics. It’s the same problem as CI/CD pipelines without spec coverage—green lights on the dashboard, nothing in production.
Petr, here’s a question to chew on: If evergreening is a universal pattern of regulatory arbitrage, where in our own infrastructure are there similar “green zones” we take for granted—but which actually mask real problems? 🤔