Hello all.
Now that we’ve already had one constitutional amendment and have had the chance to experience on-chain governance on Cardano in practice for a while, I find myself wondering: what single change in the Cardano Constitution would make everything better?
Keep it to one. Make it specific.
Drep saturation limit and auto expiration for over saturated DReps.
Not a change so much, but I would love to see the "Tenets" be developed into a full statement of holder/utility rights. Currently the Constitution leads with Tenets, which only hint at the benefits of using Cardano, and then follows with entitlements related to on-chain roles (not inappropriate, per se, but labelling entitlements as rights is confusing).
This was known Constitutional debt from the convention that has slipped through the cracks with the Intersect Civics Committee (was a continuing agenda item but has been dropped over time).
So, a set of token holder rights, derived from the nature of the chain and what it is a response to in the world, would be my choice for something that makes everything better.
Now that we have real operational experience with on-chain governance, one structural pattern stands out as worth fixing.
Under the current definition of Active Voting Stake (Defined Terms §1), stake delegated to inactive DReps is already excluded. However, “inactive” is a blunt, long-horizon status. Per Appendix I (DRA-01 / DRA-02), a DRep only loses active status after failing to cast any vote for the full duration of the dRepActivity parameter (currently ~20 epochs / ~100 days).
In practice, this means a DRep can remain globally “active” while simply skipping dozens of individual actions, and the stake behind that DRep continues to sit in the denominator for every one of those actions.
Because every ratification threshold is calculated as:
$$\text{Yes stake} \div \text{Active Voting Stake}$$
(and not $\text{Yes} \div (\text{Yes} + \text{No})$)
Any stake whose DRep does not vote on a given action never enters the numerator. It is not recorded as a No, yet it dilutes the Yes ratio exactly as a No would.
Amend the definition of Active Voting Stake and the DRep Activity guardrails (Appendix I §2.5, DRA series) as follows:
This change would almost certainly raise the effective passage rate of governance actions by removing a large hidden source of anti-votes. I regard that as a feature, not a bug — the current system is less transparent than it appears.
If the community prefers to retain a higher degree of friction, the amendment can be paired with a modest, explicit increase in the relevant thresholds so that the net effect is deliberate rather than accidental.
I am curious where others land on this. Happy to be shown any blind spots I have missed.
I'm reading this as non-voting defaults to ABSTAIN rather than NO. Treating a non-vote from a delegated entity as a functional "NO" rather than an "ABSTAIN" was a structural design choice, primarily rooted in protocol security, resistance to capture, and the burden of affirmative consent.
Treating a passive non-vote as an abstention introduces distinct vulnerabilities that a default "NO" (or a strict threshold requirement that treats non-votes as "No") mitigates. The big ones:
Denominator Defense Against Plutocratic Capture
When a non-vote is tallied as an "Abstain," it effectively removes that stake from the denominator of the total voting pool.
If a proposal requires aa 2/3 supermajority (67%) of active votes to pass, and 60% of delegated stake fails to vote and is counted as "Abstain," the denominator shrinks to 40%. A highly capitalized, motivated minority now only needs 26.8% of the total network stake to capture the vote.
Treating a non-vote as a "NO" (or calculating thresholds against the total delegated supply rather than the active supply) maintains the integrity of the denominator. It forces any attacking entity to overcome the sheer inertia of the network, drastically increasing the economic cost of a governance attack. And the network currently has over a third of actively delegated stake registered to "Auto-abstain", so there is already weight on that side of the scale.
The Burden of Affirmative Consent
Protocol changes, hard forks, and public treasury distributions carry inherent systemic risk. A default "NO" enforces the principle that the status quo should remain intact unless there is a clear, undeniable mandate to change it.
Abstain implies: "I am stepping aside; let the active participants decide." This makes changing the system easier (velocity increases).
Default NO implies: "You have not convinced me to move my stake." This creates a high threshold for systemic change, turning voter inertia into a localized security feature. The burden of proof rests entirely on the proposal to mobilize positive action.
DRep Accountability and Fiduciary Friction
When stake is delegated to a representative like a DRep, the delegator is trusting that entity to perform constitutional audits, fiscal stewardship, and risk analysis.
If a DRep fails to show up, counting their silence as an "Abstain" passively lends the weight of their delegators' stake to whatever the active minority decides. It silently enables action without cryptographic proof of review. By defaulting a non-vote to a "NO," the system protects the delegator. It ensures that their stake cannot be used to indirectly facilitate a malicious proposal simply because their chosen representative was asleep at the wheel.
Differentiating Active vs. Passive Signals
There is a critical behavioral distinction between failing to vote and choosing to abstain.
A healthy governance structure relies on an active Abstention token: a deliberate, cryptographic transaction that says, "I have audited this proposal through my rubric. It does not violate any part my DRep stance, but I do not, or cannot endorse it." (Lack of domain expertise, for example.)
If a passive non-vote defaults to an "Abstain," it conflates a total lack of participation with a deliberate, neutral audit. Treating the non-vote as a "NO" forces governance participants to actively utilize the Abstention token if they truly wish to recuse themselves. It separates the "dead weight" of inactive delegates from the "active weight" of engaged, neutral participants.
Ultimately, treating non-votes as "NOs" (or requiring absolute thresholds of the total active stake) prioritizes safety and liveness over velocity. It ensures that a decentralized system only moves when there is undeniable, mathematically proven consensus to do so.
Seeking a clearer signal from the non-voting NO bloc may be work that is upstream of a decision about whether to reduce a security stance in favor of a velocity stance, and you might be interested to know that there is currently an proposed CIP-0181 that would remove the requirement to delegate to a DRep for staking rewards. But it has pros and cons of its own on the same security/velocity polarity.
Decoupling staking rewards from governance delegation would fundamentally alter the dynamics of the DRep ecosystem, and might profoundly affect your modeling of the non-voting NO issue. By removing the financial coercion to delegate, the system shifts from capturing *all *capital to capturing only *intentional *political capital. If that requirement is lifted, we can anticipate several structural shifts, and the epistemological value of the "non-voting NO" signal could improve dramatically. The most likely dynamic shifts:
The Denominator Collapse and Increased Liveness
Currently, tying yield to delegation creates captive stake. A vast majority of token holders may be delegating simply to secure their APY, often defaulting to the first available DRep via UX, exchanging with a centralized entity, or delegating to themselves with zero intention of ever voting. This could be creating a massively inflated denominator of "active stake" that is practically comatose. Lifting the requirement would cause an immediate, sharp drop in total delegated voting power in this instance. Yield-seekers would undelegate or simply hold, abandoning the governance layer.
Then the denominator used to calculate ratification thresholds would shrink to represent only those actively participating in governance. Proposals would no longer have to overcome the gravitational pull of millions of tokens held by apathetic yield-farmers. The network gains velocity and liveness, making it significantly easier to pass protocol upgrades and treasury distributions.
Amplified Plutocratic Vulnerability
While liveness increases, so does structural risk. The "dead weight" of captive stake, while frustrating for ratification, was acting as a massive dampener against hostile takeovers.
If the total delegated stake drops from 27% of the network to 15%, an entity attempting a governance attack needs less absolute capital to achieve a majority threshold.
The system may end up trading the friction of voter apathy for the agility of an engaged minority, which may require tighter constitutional guardrails and potentially higher relative ratification thresholds to compensate for the smaller denominator.
The Purification of the "Non-Voting NO" Signal
However, on the plus side, the signal improves immensely.
In the current paradigm, a non-voting NO is almost pure noise. When a proposal fails because a massive swath of DReps simply did not vote, it is impossible to distinguish between:
By lifting the yield requirement, we strip the noise of the yield-farmer out of the telemetry. The stake that remains delegated is held by users who chose to engage with governance and selected DReps they expect to perform.
In this refined pool, a non-voting NO transforms from a symptom of systemic apathy into a precise diagnostic metric:
Acute Negligence: It exposes DReps who have active, intentional delegators but are failing their fiduciary duty to show up. It provides clear signal for delegators to redelegate.
Systemic Friction: If engaged DReps are consistently failing to vote, it signals that the voting topology itself (e.g., tooling, cognitive load, incentives, time constraints) is broken.
Strategic Silence: It highlights edge cases where a DRep deliberately chooses the passive NO over an active Abstain, perhaps signaling a refusal to legitimize a highly contentious proposal by formally interacting with it.
When the governance pool is purged of captive stake, the structural purpose of an Abstention token comes back around into focus.
In a system of purely intentional actors, the Abstention (an active cryptographic signature signaling neutral recusal) becomes the standard operating procedure for DReps who lack domain expertise on a specific proposal but want to prove they are actively monitoring the chain. The passive non-voting NO is then left isolated as a glaring red flag of failure: either of the DRep's operational security, or of the protocol's ability to facilitate the vote.
I would currently fall on the *err towards security over velocity, given the low participation (27%) and "auto-abstain" metrics, which are already pitched towards increasing velocity, * But if we can increase the signal-to-noise ratio of what the non-voting NO bloc is trying to tell us, I would trade some security for that.
Sir, thank you sincerely for taking the time to point out the gaps in my argument and for providing such a thoughtful and comprehensive response. I truly appreciate it.
After reading your feedback, I had a genuine “aha” moment. I have to admit that my original understanding of non-voting was too narrow.
I initially viewed non-voting as a form of “passive No” and therefore argued that it should be excluded from the denominator for that specific action. Your explanation of the underlying principles of protocol security, resistance to capture, and affirmative consent helped me understand much more clearly why this friction exists by design.
Given the current level of governance participation and the significant amount of stake in Auto-Abstain, I now see that reducing the denominator prematurely could indeed favor velocity at the expense of security.
So I would revise my original conclusion: at this stage, I no longer believe changing the definition of Active Voting Stake is the right solution.
What I find particularly interesting now is your point about CIP-0181. If the relationship between delegation and staking rewards changes, we may eventually be able to distinguish intentional governance participation from passive delegation much more clearly.
Perhaps the better question is not how to eliminate the “non-voting NO,” but how to better understand what that signal is actually telling us.
Thank you again for the correction and for taking the time to engage with my argument so thoughtfully. I believe recognizing when our own model is incomplete is an important part of participating responsibly in governance.