Feedback Request: a new complementary maximal bound issuance for the Quebec protocol proposal

[NomadicLabs]

This is a joint post by Nomadic Labs, Trili.tech and Functori

The Quebec protocol upgrade proposal is almost ready! But before announcing the proposal, we want to reach out to bakers to gather their feedback on one particular change in the Adaptive Issuance mechanism: the introduction of a new adaptive maximum.

In a nutshell, this new bound makes maximum issuance dependent on the staked ratio. It prevents potential future scenarios where a stagnation of global staked tez close to the target ratio could lead to undesirably high issuance rates.

In order to smooth out the governance process and the potential adoption of Quebec, we want to make sure that the current implementation of the proposed new feature has sufficient consensus within the Tezos community. Hence, this post provides both a higher-level description of adaptive maximum and an opportunity to experiment with the chosen definition.

Adaptive maximum in the Quebec proposal

Issuance rates on Tezos mainnet are currently bounded by progressive maximum and minimum rates. They determine the acceptable limits for the issuance rate as a function of the time elapsed since the activation of Adaptive Issuance – i.e. from cycle 748. From the start of cycle 808, not expected before December 8th, the progressive maximum bound will remain constant with a value of 10%. This could lead to scenarios where issuance rate can grow undesirably high, even when the staked ratio is close to the 50% target.

The Quebec proposal addresses this issue by incorporating a complementary adaptive maximum that decreases smoothly when the distance to the target staked ratio decreases. As shown on the plot below, the closer we are to the target staked ratio, the lower the adaptive maximum bound is:

Please refer to Quebec’s documentation for further technical detail, including the specification for the chosen curve and its interaction with the final adaptive issuance rate. The following table illustrates selected data points.

Staked ratio	5%	10%	20%	30%	40%	50%
Adaptive maximum issuance (approx.)	10%	8.4%	3.9%	2%	1.2%	1%

Regardless of the value of the new adaptive maximum bound, the final issuance rate will never be below the progressive minimum rate. This prevents issuance from dropping to undesirably low rates.

We refer readers to the Quebec proposal’s documentation for further technical detail on this feature, including the specification of the chosen curve and its interaction with the final adaptive issuance rate.

Call for action

As we mentioned above, we would like Tezos bakers and the community at large to provide their input and feedback on the proposed definition of the adaptive maximum curve, before we announce that Quebec is ready.

The following notebook can be used to experiment with the definition of the adaptive maximum, and plot the resulting curve.

AdaptiveMaximum.ipynb

We invite all interested community members to provide their feedback in this thread, including proposals for alternative definitions. We will assess this feedback before formally releasing the Quebec protocol proposal.

Thank you in advance!

[Primate411]

I like this proposed change. With this in place Adaptive Issuance (AI) starts to make much more sense as a way to minimize emissions while preserving an adequate level of protocol security.

In addition to the progressive minimum rate, we should seriously look at the staker/delegator reward split. If protocol security is what we seek ultimately, rewarding delegators with 50% of the staking rewards doesn’t make sense, especially in the medium-to-long term, if IA is to accomplish its goal.

I think a good place to start is reducing delegator rewards by 33% and putting them into staking rewards.

[LiberteZ]

+1

I’d suggest reducing Delegator rewards to 1/10 of what Stakers receive, or even eliminating them entirely like other protocols. This would create better alignment in terms of incentives, especially for yield optimization and L2 incentives. Right now, many Delegators are content to stay at L1, earning 7% with zero risk, instead of staking or exploring L2 options for higher yields.

[Primate411]

Got a position from PosDog Telegram: Contact @posdog

[TEZ_Blocks]

I don’t oppose this at all, it’s a great addition to the Adaptive Issuance feature.

However, as of now i feel like there is low incentive to change from delegating to staking.

And that is because as the issuance of new coins rises, so do delegator rewards.
Even now with Adaptive Issuance in full swing delegators are earning more than before the upgrade.

IMO delegator rewards should remain half (or even less) of what stakers earn untill they reach 5% (max issuance for delegators). Staker rewards should keep rising within the necessary bounds.

[BakeTzForMe]

I don’t understand the calls to further lower the delegation rewards on the protocol level. You are fully in control of how much your delegators are earning. If you think your delegators are earning too much, then reduce how much you pay out to them, or stop paying them entirely. That will provide them with an incentive to do something else with their funds.

[Cryptonio]

While you have a point, I think if something like that is not implemented on the protocol level, all you will achieve by raising your fees is to damage your competitiveness.

@BakeTzForMe

That will provide them with an incentive to do something else with their funds

Most probably to redelegate to another baker with lower fees.

[LiberteZ]

Exactly this.

[BakeTzForMe]

@Cryptonio

I think if something like that is not implemented on the protocol level, all you will achieve by raising your fees is to damage your competitiveness.

Implementing something like that on the protocol level damages my competitiveness! If delegation rewards are reduced even further then I will have to raise my fees in order to stay afloat.

AI already cut my rewards in half, whether by reducing delegation rewards to 50% or by limiting the maximum external stake to 5x my own stake instead of the 9-10x it is for delegation rewards.

Why should the protocol force me to sacrifice my competitiveness when instead it can give you the option of choosing whatever levels of competitiveness are acceptable to you?

If you really feel delegators need to earn less rewards to incentivize different behavior, then put your money where your mouth is.

@Cryptonio

Most probably to redelegate to another baker with lower fees.

Yes, but with increased demand comes decreased supply. Eventually the bakers with lower fees will become full or over-delegated, at which point it will make sense for them to also raise their fees to strike a balance between supply and demand. In either case (over-delegation or higher fees), it will no longer make sense for additional delegators to delegate to those bakers, and they’ll be incentivized again to delegating elsewhere or to start staking.

The rising tide (delegation fees) lifts all boats (bakers / staking).

[Boulange]

We fully support this proposal.

I was actually concerned about the possibility of having a high issuance rate even when the network is sufficiently secure (or close to), which is contrary to the AdIs philosophy (AI means something else to most people ).

[TEZ_Blocks]

Blockquote “Implementing something like that on the protocol level damages my competitiveness! If delegation rewards are reduced even further then I will have to raise my fees in order to stay afloat.”

If you implement it on protocol level it will be equal for all bakers, so it can’t hurt your competitiveness.
Further more it will increase the chance that a delegator will start staking instead of redelegating to someone else.

Blockquote "AI already cut my rewards in half”

That can only be true if you don’t accept stakers.

The max X5 limit for stakers, equals the x10 limit for delegators. So that can’t hurt you
Further more, now after the upgrade you can stake your entire stack instead of having the amount of delegators dictate that.

So you SHOULD be earning more than before.

What cryptonio says is true, if a baker raises their fees, they will lose delegators to a different baker. That is self sabotage

[tokyo_on_rails]

While reducing the inflation rate was one of the original goals of AI and this aligns with that, this design seems to be skewed in favor of the lower issuance rates at the risk of losing the incentive to stake.

The issuance rate evens off dramatically between 30-50%. If 30% of the network is staked then the maximum will be only 2%, and the actual adaptive issuance presumably often lower than that.

If 50% staked is still the target, I struggle to imagine bakers/stakers choosing to lock funds and stake for such a small difference in rewards, when we’re already having trouble getting people to stake when they are earning 15% on a 10% max. In line with this, I also agree that we could increase the ratio that stakers receive vs. delegators to help disincentivize delegation.

If the goal is simply to reduce inflation then this is ok, but if we really do want 50% of the network staked, i see little incentive here to continue staking until 50% vs. staying liquid if the rewards are negligible.

Maybe the curve could be more straight like this (red):

[BakeTzForMe]

@TEZ_Blocks

@BakeTzForMe

AI already cut my rewards in half

That can only be true if you don’t accept stakers.
The max X5 limit for stakers, equals the x10 limit for delegators. So that can’t hurt you

Okay, so first of all, I realize now that I’ve been careless in the words I’ve used, and I apologize for that. I used the word “rewards” when what I was talking about was “baking power” instead. This is an old habit from before Paris when rewards were directly correlated to baking power and basically nothing else.

I immediately staked as much of my balance as I could and configured my baker to accept 5x external stake, but my baking power still dropped to less than 60% of what it was prior to Paris. That’s not as bad as half–meaning I made some broad generalizations and rounding–but it’s still much worse than I’d like!

Now let’s examine your statement that the new values in Paris can’t hurt me.

Before Paris:

• Delegation was worth 100% baking power and could be 10x of your balance. 10x100% == 10x

Starting with Paris:

• Staking is worth 100% and can be up to 5x of your staked balance. 5x100% == 5x
• Delegation is worth 50% and can be up to 10x of your balance. 10x50% == 5x

So yes, the limits of staking and delegation are equal to each other in terms of baking power. But that’s still half (each) of what it was before Paris. The only way to have the same maximum baking power you had before Paris is to be at the limits of both staking and delegation.

This means starting with Paris you need to retain all of your delegated balance and gain an additional 50% as much staking balance as you have delegated balance in order to break even in terms of baking power.

If you eliminate the value (in terms of baking power) of delegations in an effort to chase more staking you will only attain up to 50% of the max baking power you had prior to Paris. You’re robbing your left hand to pay your right hand.

@TEZ_Blocks

Further more, now after the upgrade you can stake your entire stack instead of having the amount of delegators dictate that.

I don’t understand that.

Before Paris:

• My “entire stack” was always worth 100% baking power regardless of delegations.
  • This included liquid funds placed in a hot wallet to pay delegators.

Starting with Paris:

• Only my staked balance is worth 100% baking power.
• My liquid funds placed in a hot wallet for payouts are worth only 50% baking power.
• Any leftover rewards earned from delegations are worth only 50% baking power unless I manually stake them.

@TEZ_Blocks

So you SHOULD be earning more than before.

As shown above, my baking power is in fact reduced starting with Paris. The only reason I am earning more rewards right now is because the rate of issuance went up due to low overall staking during this transition period. While the current increase in rewards is nice, I see it as temporary because I believe that Adaptive Issuance will achieve its long-term goal, especially as it fully goes into effect over the next few months. Meaning I believe staking will reach the ~50% target and issuance will go down to ~1% at most (when this proposal for Quebec goes through).

Hopefully I’ll get more externally staked funds by then to help bring my baking power back up to pre-Paris levels. But if the delegations disappear then I’ll still only have at most 50% of the baking power I could have had before Paris went into effect.

@TEZ_Blocks

If you implement it on protocol level it will be equal for all bakers

What cryptonio says is true, if a baker raises their fees, they will lose delegators to a different baker. That is self sabotage

I agree with you that what you want is self-sabotage. What you seem to be saying is that you will only go through with it if you can force everyone to be sabotaged equally. What I am saying is don’t force it on me!

[Val]

Dear Community

this is my only statement on this matter. I do not have time to discuss it further, and this statement is solely in response to the original proposal, not the ongoing discussion.

We have consistently emphasized that adaptive issuance is fundamentally about ensuring network security, and we’ve guided the community toward understanding this principle. However, the current proposal regarding the adaptive maximum is fundamentally flawed and risks undermining the very security that adaptive issuance is designed to protect.

Adaptive issuance is supported by extensive research, not only within our framework but also in other projects like Polkadot and similar networks. I have not encountered any precedent for this kind of adjustment, which raises concerns. This absence of precedent does not automatically render the proposal incorrect, but it does highlight the lack of supporting research or evidence directly linked to this change. Any modification to the economic protocol should be thoroughly backed by evidence-based research.

Allow me to explain why this proposal is problematic. The core principle of adaptive issuance is that the network remains secure when 50% of the total supply is staked. The current proposal suggests capping maximum issuance on the assumption that the chain is sufficiently secure. However, compromising on security is a misguided approach. The intention behind reducing issuance is to avoid overpaying for security, but this rationale is fundamentally flawed. Currently, we are not even able to reach the secure state, even with higher issuance rates, which means we are not overpaying; in fact, we are still underpaying. Yet, despite this, we are seeing a proposal to reduce emissions.

Moreover, introducing additional complexity into an already intricate system without robust economic research or simulations is unwise. The current proposal seems like an attempt to resolve market issues with another engineering solution, rather than addressing the underlying economic factors. A simpler and more effective approach already exists—driving adoption and improving market performance by delivering value, i.e., creating something useful.

There are better alternatives to consider. If we believe the chain is secure with a lower staking ratio, we could adjust the target or broaden the acceptable range. If the goal is to halt emissions, we should consider stopping issuance entirely and relying on voluntary stakers.

It’s important to note that there is a valid reason for maintaining a high issuance limit, even when the staking ratio is close to the target. Adaptive issuance is designed to secure Layer 1 (L1) in an era dominated by rollups and Layer 2 (L2) solutions by ensuring that at least 50% of the total supply is staked. If the staking ratio falls short, emissions are increased to incentivize the movement of funds back to L1. This shift can occur either through stakeholders moving their assets to L1 or through emissions, which dilute the balances on rollups, thereby favoring L1. By reducing emissions at higher staking levels that are still not close to the target, we risk compromising L1’s ability to recover to the 50% stake threshold, especially when rollups hold funds tightly.

In conclusion, economic protocols should not be altered prematurely, especially before they have had the opportunity to demonstrate their effectiveness.

P.S. If the goal is to balance delegation rewards and further encourage staking, consider introducing a delegation reward curve. This curve could decrease delegation rewards when the staking ratio is low, thereby incentivizing more individuals to stake directly. Conversely, as the staking ratio increases, the rewards could be adjusted upward to maintain equilibrium. This approach would more effectively align incentives with the desired staking outcomes.

With love,
V

[LiberteZ]

I find this approach very sensible. Waiting for AI to stabilize is a wise move, but lowering Delegator rewards should be tackled immediately, imo.

As time goes on, we’ll gather more data on AI, allowing us to better assess the optimal long-term solution.

[Primate411]

An assumption we have to tackle for sure is whether 50% staked is an ideal target to strive for. Based on that, the new maximum inflation numbers proposed here might be seen in a new light.

Q: What is the minimum protocol security level? (7% has “worked” previously)
Q: What’s the ideal protocol security level? (50%?)

[cryptEMES]

Tezberry Pie supports the proposal.
I have a related question though. What is the reason why a baker can only accept 5x stake? Could we raise this limit?
The before mentioned topics of lowering delegation reward in protocol and the worries of bakers earning enough to be operable could be countered with small bakers accepting more stake. Also the distribution of voting-power could shift more in favor of small bakers and so running a baker make more attractive → supporting decentralisation.
Could this be included in Quebec?

[Huge_Holder_Tez]

This proposal looks like the development team has a lot earned tokens for last 6 yesrs and wants to infringe on the rights of holders regarding the rewards they receive for staking as much as possible. Simply terrible behavior that would be called in financial circles deceiving investors and bondholders. Thus, the company would be marked and would never receive additional funding and recognition among investors. It is truly a disgrace to change conditions on the fly by a limited circle of people without taking into account the opinions of real token holders. Good luck in this circus

[Boulange]

The original post explicitly states:

@NomadicLabs

In order to smooth out the governance process and the potential adoption of Quebec, we want to make sure that the current implementation of the proposed new feature has sufficient consensus within the Tezos community.

Core teams encouraged bakers and stakers (holders) through social networks to participate in this public discussion with arguments. Then, should the proposal be submitted, it will be subject to a vote by bakers, whose interests include:

• improving the tokenomics,
• satisfying and attracting holders through their vote.

So it looks like holders’ arguments are highly valued.

[cryptEMES]

What is holding any „real token holder“ back from participating in the process of making decision like this?