The Coalition Tax

Why does the obviously better pricing model never win? And why does every banking reform quietly drift back toward the biggest banks? Same answer. It is sixty years old, and almost no strategist has read the man who proved it. The obviously better idea loses on purpose. A small organised group beats a large unorganised one even when the large one is right. That is a law, not a tendency, and it has held for sixty years.

We tell ourselves a comfortable story. Good ideas win. Markets are efficient over time. Bad pricing, bad rules, bad structures get competed away. Reform is just a matter of evidence and patience. It is not. The man who proved why was an economist most strategists have never read.

Olson's Arithmetic

Mancur Olson, in The Logic of Collective Action (1965) and The Rise and Decline of Nations (1982), demolished the comfortable story. His result is almost arithmetic. A small group with a concentrated interest reliably beats a large group with a diffuse interest. The few are organised because each member gains a lot. The many stay passive because each member loses only a little. Concentration of benefit beats breadth of harm. Every time.

Olson went further in 1982. The longer a society stays stable, the more it gets colonised by what he called distributional coalitions: organised groups built to redistribute existing wealth toward themselves rather than create new wealth. His evidence was the post-war record. Germany and Japan grew fastest because external shock had wiped their coalition density to zero. Britain, the oldest stable democracy and the most coalition-encrusted, grew slowest. Stability is not free. It accrues a tax.

Give that tax a name. Once you see it you cannot unsee it, and it sits across both of my worlds: professional services and capital markets. Call it the Coalition Tax: the gap between the system we would design and the system we get, priced by whoever is organised enough to defend the status quo.

The Billable Hour

Every serious person in professional services has known for decades that the billable hour is a poor instrument. It rewards effort over outcome. It punishes efficiency. It misaligns the firm and the client at the exact moment AI makes the hour nearly worthless. Value-based pricing is not a new idea. It has been the obviously better idea since the 1990s. So why has it stayed marginal for a generation? Not because it is wrong. Because it is an Olson problem wearing a pricing label.

Run the logic. Value pricing concentrates a large, sharp cost on a few senior billers, the partners whose pay, status, and leverage are built on billed time. It spreads a small, diffuse benefit across many clients, none of whom alone will fund the fight to change it. Concentrated cost, organised defenders. Diffuse benefit, unorganised beneficiaries. Olson tells you who wins. He has been right for sixty years.

The billable hour is not an accounting convention. It is a distributional coalition artefact. It survives reform after reform because the better model asks the organised few to vote against themselves.

The Rules of the Market

Move to capital markets, where I spend my board-level hours. The pattern is not weaker. It is sharper, because the stakes are systemic. Post-crisis rules are written, in large part, by the same coalition whose rent-seeking helped cause the crisis. The people most expert in the plumbing have the most concentrated interest in how the plumbing is rebuilt. They show up. They draft. They negotiate the technical annexes nobody else reads. The public interest is real, enormous, and diffuse. Olson predicts the diffuse interest loses the fine print.

Look at the Basel III endgame. The reform began as a hard tightening of bank capital. Watch the successive revisions and you see the gradient: each round drifting back toward the preferences of the largest dealer banks, softened, recalibrated, phased. The headline survived. The bite migrated. That is not a conspiracy. It is concentration of interest doing exactly what Olson said it would do to a diffuse counterparty.

Then look one layer down, at the market structure I work inside. Central counterparty membership rules. Clearing fund waterfalls. The exact ordering of who pays, and how much, when a clearing member defaults. We are taught to read these as pure efficiency: the optimal engineering of systemic risk. Read them as Olson and a second picture appears. High membership thresholds concentrate access among large incumbents. Waterfall structures distribute loss in ways that quietly favour the organised members who helped design them. Some of that is genuine prudence. Some of it is a coalition artefact dressed as a risk model. The skill is telling the two apart. Most market participants never ask the question.

A market inefficiency that survives reform after reform is rarely irrationality. It is usually a stable equilibrium with concentrated winners and diffuse losers. Olson's mathematics, hiding in your rulebook.

The Live One: AI Governance

Now watch the coalitions forming around frontier AI in real time. This is the moment the script runs again, and the architecture will set before anyone outside can reset it. The labs are writing their own safety frameworks. The large platforms are setting the API and tool standards everyone else must build against. The hyperscalers are positioning to run the certification and evaluation bodies. Each is a concentrated interest organising early, while the public interest in how AI is governed stays vast, genuine, and completely unorganised.

That is the Olson setup in its purest form. His deepest finding is about timing. Coalitions harden during stability and only reset under external disruption. The window in which an outsider can shape the foundational architecture of AI governance is open now and closing fast. Once the standards calcify, the Coalition Tax on AI gets baked into the substrate, and we pay it on every transaction for a generation, without ever seeing the line item.

The Steelman, and the Test

The steelman is stronger than the cynics admit. Sometimes the organised few are also correct. The senior partners defending considered pricing may genuinely understand delivery risk the client does not. The dealer banks negotiating Basel may genuinely understand contagion the public never sees. The labs writing safety frameworks may, right now, be the only people who understand the failure modes. Expertise concentrates for the same reason interest does, and the two are often found in the same room.

Coalitions are not only extractive. They are also ballast. A system with zero coalition density is not a utopia of pure efficiency. It is Germany in 1946: reformable precisely because it has been shattered. Coalitions protect stability against churn. They are institutional memory with skin in the game. Sweep them away and you do not get the system you would design. You get whoever grabs the rubble fastest.

So the test is not whether a coalition exists. Coalitions always exist. The test is whether this one is still paying for itself, or just collecting. That is the whole discipline. Not abolishing the tax. Pricing it honestly, and asking who set the rate. You will never design the system from the outside. The organised few will always draft the fine print. But you can refuse to mistake their interest for your efficiency, and you can move in the narrow window before the structure sets.

The reformer's real question was never whether the idea is better. It always is. The question is who is organised enough to make sure it loses, and what it will cost to outlast them. The status quo is never free. Someone is always collecting the Coalition Tax. The only choice you get is whether you can see the invoice.