Back to: Certified Decentralization Analyst
Governance answers the question that infrastructure and capital can’t: who can change the rules, and how hard is it? Every chain eventually needs an upgrade, a fix, a fork. Who gets to decide is a pure test of whether “decentralized” survives contact with reality.
On the surface, it looks simple
– On-chain governance — token-holders vote; the outcome executes in code (Cosmos, Arbitrum, Polkadot). – Off-chain — a foundation, core team, or few signers decide (Bitcoin’s rough consensus + BIPs, many L1s in practice).
What we actually measure
– Vulnerability / veto power — can one multisig or one account change the rules or pause the chain? (Arbitrum’s 9/12 Security Council is exactly this; Bitcoin has no such switch.) – Upgrade mechanism — do upgrades require broad validator/community consensus (node opt-in) or a foundation push? – Proposal & voting quality — participation rate, voter diversity, and whether votes even bind.
Reading the scores
| Score | Reading | |
|---|---|---|
| Bitcoin | 95/100 | No CEO, no kill switch; change requires rough consensus across a huge, adversarial community |
| Ethereum | 80/100 | Node opt-in upgrades; heavy influence from the core dev ecosystem + staking whales in some debates |
| Arbitrum | 68/100 | Fully on-chain DAO with binding execution — but a multisig Security Council holds emergency power |
The analyst's two traps
1. “On-chain voting” ≠ decentralized. If 90% of votes come from 3 wallets or a foundation, on-chain is decoration. 2. “No governance” ≠ no governance. Every chain has power centers: core teams, funded foundations, key maintainers. Find them — they may be permanent institutions that outlive “governance” decks.
Ask: “If I disagree with a rule change, who do I fight?” The answer — a
> community, a committee, or a company — is the truth of governance.
Next lesson: the Software pillar — the single-client trap.