Token governance: voting power is not ownership of cash flows

Identify the actual right

A governance token can grant influence over parameters, upgrades or treasury actions without granting a legal claim to company earnings. Its rights come from the particular contracts and governance arrangements, not from the word governance. OpenZeppelin’s Governor documentation shows how proposal, voting and execution components can be assembled. The flexibility means analysts must inspect the configuration rather than assume a standard constitution.

Start with three separate questions: who may propose, whose votes count and who can execute a successful proposal? A fourth question is whether another authority can veto, pause or override it. The answers can differ, and a token holder may control only one small part of the process.

Measure effective participation

Total token supply is not the same as actively delegated voting power. Some tokens may sit in contracts, remain undelegated or be held by participants who rarely vote. A small percentage of total supply can dominate the votes actually cast. Conversely, a quorum measured against a broad supply denominator can make passage difficult even when most participants agree.

Imagine 100 million outstanding tokens, 20 million delegated and 6 million votes participating in a proposal. An entity controlling 4 million voting units owns only 4% of total supply but supplies two-thirds of that vote. This does not prove improper conduct; it shows why concentration should be measured against both supply and participation. Explain the denominator every time.

Follow a proposal to its effects

A forum post, a signaling poll and an executable transaction are not interchangeable. Read the proposed contract calls and parameter changes, not just the headline. Determine when voting power is measured, how the voting period works and whether execution occurs through a timelock. A delay can create time for review, but it does not establish that every affected user has an available exit.

For a paper review, translate a proposal into a before-and-after table. If a borrowing threshold changes, identify which positions might become vulnerable. If a fee changes, identify who pays and who receives the difference. If funds leave a treasury, distinguish a grant, an investment and an irreversible transfer. These economic effects matter more than the popularity of the proposal’s title.

Separate treasury assets from token value

A treasury balance does not automatically belong proportionally to every token holder. Spending authority, legal arrangements and redemption rights may prevent that interpretation. A treasury consisting largely of the protocol’s own token is also different from external liquid assets: selling it can change both price and governance distribution.

Suppose a project reports a $50 million treasury, of which $40 million is its own thinly traded token at the last price. Calling the project cash-rich obscures the $10 million of other assets and the uncertain realizable value of the rest. A useful analysis separates liquid operating resources, restricted balances and self-issued holdings rather than adding them into one apparently cash-like number.

Evaluate changing control over time

Token unlocks, delegation changes and low turnout can alter effective control without any change in the core software. Review historical participation, major delegates and upcoming supply changes. A dispersed holder list is insufficient if voting power is delegated to a few actors; a concentrated list is not the whole story if those holders face enforceable restrictions.

Governance quality is difficult to compress into a single score. Document procedural safeguards, conflicts of interest, execution history and unresolved uncertainties. A well-run process can still approve a financially poor decision, while a technically sound proposal can fail through poor participation. Treat governance as an evolving source of operational and economic risk, not a badge that guarantees decentralization or future token appreciation.

For a governance review, distinguish a rule that exists in code from a convention that participants usually follow. A community may normally discuss proposals for a week even though the contract permits a faster process. Record both the customary process and the enforceable minimum. Stress analysis should use the latter when assessing how quickly control or economic parameters could change.

A large percentage can describe a small electorate

Suppose a fictional governance token has one million eligible votes. A proposal attracts 50,000 votes, of which 35,000 support it. The approval rate among participants is 70%, but supporting votes represent only 3.5% of the eligible total. If the hypothetical rules require 40,000 participating votes for quorum and a simple majority of those voting, the proposal passes. Different quorum and counting rules could produce a different result from the same ballots.

Write the denominator every time you report a percentage: all issued tokens, eligible voting power, delegated votes, participating votes or votes in favor. These are not interchangeable. An impressive mandate can shrink when the denominator changes. Equally, low participation alone does not prove a proposal is illegitimate; you still need to understand the rules and the distribution of effective control.

Repeat the example with 30,000 of the supporting votes controlled by one delegate. Then ask how independent the remaining voters are and whether delegators can withdraw that authority in time. The exercise shifts attention from the number of addresses to the people or entities that can coordinate a decision. Where their independence cannot be established, leave that uncertainty visible.

Follow the vote all the way to execution

A passed proposal is not always an executed change. Governance systems may include a queue, a delay, cancellation powers and a separate executor. OpenZeppelin’s governance components illustrate how these stages and roles can be configured. Read its governance documentation. The configuration of the deployed system matters more than the generic label “DAO.”

For a fictional proposal to change a collateral limit, build a timeline: voting begins, the snapshot is fixed, voting ends, the action is queued and the change becomes executable. Mark who can stop or alter the process at each point. If an emergency administrator can bypass the ordinary delay, include that route rather than presenting the standard vote as the only form of control.

A delay can give observers time to notice a change, but notice is useful only if they can act. If withdrawals take longer than the delay or a separate control can pause them, the practical exit window is different from the advertised governance window. Evaluating governance means connecting decision rights to the consequences for users, not merely checking whether voting occurs on-chain.