Tokenomic Alignment with L1 Growth and Architecture #246
Replies: 3 comments 2 replies
-
|
Regarding (3) above: Note that the for the 0.000001% of total C-Chain Stake weight multiplier for fee basis, although seems to be a good approximation, this could be further refined from a linear multiplier to something that more formulaically approximates the marginal benefit to base chain security/value of adding the additional validation/stake weight at scale , such that marginal increases in total stake weight have a diminishing effect at scale. An asymptotic or logarithmic function likely best approximates marginal value at scale. |
Beta Was this translation helpful? Give feedback.
-
|
As this progresses, it is likely best to focus first on a simplified version of this ACP that moves forward parts (1) and (3), or to even break those up for separate consideration also. So in its minimal prioritized form, I believe (1) (fees directed to validator/stake rewards) is the simplest and most essential starting point. |
Beta Was this translation helpful? Give feedback.
-
|
Hey, thanks for taking the time to write this up. I would like to better understand how you imagine some of the details working in practice. Regarding (1), the shift from burning fees to directing them to validator rewards, I think your logic broadly holds at the system level: if governance can dial down inflationary emissions as L1 fee flows grow, the net effect can absolutely be neutral or even deflationary. The challenge I see is less about the mechanics and more about the narrative and game-theory for non-validating AVAX holders. Today the “AVAX burn” story is extremely simple and powerful from a market perspective. Replacing a significant portion of that with rewards that end up as sellable tokens in validator hands will inevitably be perceived by some as structurally more inflationary, even if the macro accounting is more subtle. Have you thought about preserving a guaranteed, non-trivial burn floor that is harder to change via governance, and only redirecting the marginal portion of L1 fees to validators as in (1)? That could help reconcile the “sound money / deflation” crowd with your goal of tying L1 activity directly to staking rewards. Some simple numeric scenarios “before vs after” (net issuance, burn, validator APR) under different L1 volume regimes would probably help a lot to make (1) more digestible. A second concern touches both (1) and your general framing of how this evolves over time. You are implicitly assuming a trajectory where L1s grow strongly, so that their fees become an increasingly large share of validator rewards, which in turn allows emissions to be reduced. I agree this is the ideal path. But what happens in less optimistic scenarios where L1 growth is slower, more cyclical, or where a non-trivial share of important L1s end up with disjoint validator sets and low on-chain activity from the point of view of AVAX fee flow? If we have already reduced burn and redirected fee flows along the lines of (1), but L1 volume does not materialise as hoped, we could end up in an awkward situation: lower burn than today, similar or higher emissions because governance is slow to cut, and not enough L1-based rewards to compensate. Do you imagine any guardrails or thresholds here, such as only increasing the share of fees going to validators once certain L1 revenue or activity metrics are sustainably reached? On (3), the idea of changing from a static baseline fee (like 1.33 AVAX) to a fee set as a percentage of total main-chain stake weight is conceptually very elegant. L1s pay more as the security they are effectively “buying” improves, and less when the quality of security decreases. That is a clean mental model. In practice, though, there are a few potential pitfalls that might require more explicit treatment. If fee levels and baseline are tied too directly to total stake, L1 cost structures might fluctuate for reasons that have nothing to do with their own usage or behaviour, such as a large validator entering or leaving, or market cycles affecting staking behaviour. That introduces volatility that is difficult to plan around for serious builders. There is also a possible feedback loop: if fees from L1s become a big component of validator rewards and stake declines, security drops, L1s pay less by design, rewards shrink, and that can further discourage marginal validators. Finally, the more tightly fees are coupled to stake as in (3), the stronger the incentive for large stakers to push the parameters in directions that benefit them through governance. You mention hysteresis, which is the right instinct, but I would be interested in how concretely you imagine that meta-layer being implemented for (3): for example, updates only every N epochs, caps on how fast the fee baseline can move, and using moving averages of stake instead of raw snapshots. Your points on validator specialisation and L1s directing fees to a subset of main-chain validators chosen from the AVAX-staked set (and then optionally adding their own emissions on top, as described later in your text and in (5)) are, à mon sens, some of the most powerful parts of the proposal. This gives L1s a proper market mechanism to attract validators that match niche requirements in terms of uptime, hardware, geography, compliance constraints, or even “soft exclusivity” (main chain plus only a limited set of L1s). At the same time, this also opens the door to a more stratified validator landscape where well-capitalised, compliant, institutional operators systematically capture the richest reward streams and progressively accumulate more influence. That may be a necessary trade-off if Avalanche wants to be truly “institutional-grade”, but I would be curious whether you see this outcome as acceptable, or whether you think it should be counterbalanced by transparency and soft guardrails, for example via public metrics on income distribution per L1 and per validator, or some attention to concentration risks. It would be interesting to know how you view these second-order effects around (5) and its variants. On the idea you sketch about requiring the additional incentives in (5) to be provided as LP pairs (for example, native gas token plus USDC or AVAX), I think there is a lot of potential but also quite a bit of complexity. It is attractive in that it could help turn the main chain into a stronger liquidity hub and further align AVAX holders with the success of individual L1s and their tokens. On the other hand, it adds another layer of design choices for L1 teams, who must now think about pools, impermanent loss, and which pairs to use, on top of everything else. I would be interested to hear whether you see the LP-pair variant as a strong recommendation, a soft “nice to have”, or something that should be kept out of the MVP scope you mention at the end in order not to overload the initial ACP. The PAYG degressive structure you propose toward the end, where the marginal validator fee decreases as the size of an L1 validator set increases, is a very neat way to incentivise decentralisation without making large sets prohibitively expensive. It fits nicely with the rest of the design. That said, here too, I think the community will want a bit more clarity on how you picture the shape and governance of that curve. Do you imagine the form of the PAYG curve (how fast costs per validator decline, where it asymptotes) to be fixed or configurable, and what “target region” in terms of validator counts do you have in mind for more public L1s versus explicitly private ones? Since this is partly meant to change behaviour, a simple example or even just an order-of-magnitude sketch would really help others understand what (in practice) you are trying to nudge L1s toward. Stepping back from the individual points, one last aspect concerns the overall complexity of the full package formed by (1), (3), (5) and the other refinements you mention. For advanced or institutional L1s, this “menu” of options is probably ideal: dynamic fees as a function of security, routing choices between burn and validator rewards, specialisation of validator subsets, optional additional emissions in different asset forms, and a PAYG curve aware of validator set size. For many teams, though, this is a lot to absorb and reason about. It might be useful to explicitly distinguish a simple “default mode” from an “advanced mode”: a base configuration with opinionated, safe defaults where an L1 can essentially plug in and go without deep economic design work, and then a second, opt-in layer where all the knobs from (1), (3), (5) and the PAYG refinements become available for sophisticated projects. You already hint at splitting parts (1) and (3) into separate ACPs; sharpening that into a clear layering strategy would probably make it easier for the community to evaluate and iterate. |
Beta Was this translation helpful? Give feedback.
-
|
Truly excellent, thoughtful feedback / discussion. Much appreciated! I'll try to respond to some of your questions/thoughts in the coming days as time allows. |
Beta Was this translation helpful? Give feedback.
-
|
Response to ACP Comment Again, thanks for your comments! Here, I'll address just the first paragraph (of 7) of your comments/questions, regarding (1) (directing PAYG to rewards), below. Expect more to follow in later posts/discussion, to address your other comments. (Apologies for the rough draft nature of this reply. I lost all prior work when my laptop crashed and recreated it here, but am tired, so skimped on proofreading it.) First, I strongly agree that numeric examples will help to illustrate benefits and likely needed to help allay potential concerns of those whose initial impression may assume that redirection of fees from burn could be inflationary. I've played with the numbers and am confident that the net effect is highly beneficial, so I plan to present some examples with numbers (some here, but more in future posts). I also agree that these examples will be needed to help socialize the proposed change, since the dynamic is more nuanced than the simple "burn everything" story.. I think we can also demonstrate, at least anecdotally, via current data and trends (price action, validator stats, stake yield, etc.) that the status quo story isn't working (and why, and contrast it with, which likely won't take much convincing to those who have invested in $AVAX and have lived the market data. We can also appeal to intuition by laying out, in more detail, the argument that the proposed change more naturally structurally aligns the validation and token dynamics with the market and L1 growth. Another point to emphasize to those concerned about the ability to dump rewards, is that (1) calls for them to be locked for the entire staking duration of the validator. They are also distributed to all of the stakers, not just the validator, so it is much less concentrated. (Also per below, it is arguable that sell pressure also is greatly alleviated when the asset turns from a depreciating one currently, with very weak links to yield or L1 growth, into one that has growth-linked market-driven value, which I’ll attempt to show is net appreciatory. The main (and unsustainable) problem with the current dynamic with its steady long-term depreciation, is that there is already not only incentive to dump almost any staking yield immediately, but there is very little incentive (when accounting for asset depreciation) to even lock or stake to begin with, especially with low rewards coming solely from inflationary emissions. By introducing a linkage between $AVAX and L1 growth, the thesis is that this dynamic begins to change by the silent hand of the market. Note also that reducing emissions is not the primary purpose or intended effect of (1), nor is reducing emissions necessary for (1) to be net beneficial (although it likely could and would be a beneficiary effect of it, especially in the long-run). Again, to your point of addressing the perception of “losing burn”, I would point out to those with that initial impression that (1) still completely allows for its net effect on total protocol burn (or inflationary effects of emissions - burn) to remain exactly neutral (stable at current) if that is desired. In fact, it just introduces increased flexibility in governing and adjusting emissions by allowing more means to balance [emissions, rewards, fees], if desired. However, I would argue to anyone tempted to use this flexibility to try to pull levers by hand based on arbitrary targets and round numbers using the heavy hand of intervention, that the real beauty of the proposed changes are that they introduce the balancing optimizing hand of the market, which is a powerful optimization dynamic. But the bottom line is, if they were to ignore me and still want to offset the minimal inflationary effect from PAYG fee redirection by lower emissions, nothing prevents that and they in fact have more levers / less constraints at their disposal. Yet another point here to the casual $AVAX holder, who does not put the asset to work staking and would be concerned by this perceived effect. I would lay out for them that even the non-staker benefits greatly (esp. in comparison to burn) from the dynamics created by market-linked and L1-linked staking/validation rewards. As much as I wish to avoid using the term flywheel, I will lay out some examples below showing the asset valuation pressure that is net positive to them. Note too that this market dynamic of (1) is complemented by (but not dependent upon) part (3) - (fees tied to quality/security of main validation) of the proposal. While (3) adds an element of completeness to the fee and market-balanced picture, (1) can be implemented easily and independently from it, so we will leave discussing aspects of (3) to a future post. Another question we can ask the reader, to illustrate the proposal, is: “So what happens in the long-run when validator emissions effectively run out, since supply is capped, or are decreased via governance?” What alternative incentivization is there for staking/validation? Note that Bitcoin/BTC has dealt with similar questions, centered around network sustainability once block mining rewards cease and transaction fees alone incentivize network validation / security. However, in their case, transaction fees go to the validators/minors, while in the case of Avalanche / AVAX, we burn transaction fees, so even those aren’t available. The importance of incentivized validation sustainability (outside of finite and inflationary emissions) is therefore even more important for us. So what replaces those incentives to keep the validation healthy? This proposal provides that means indefinitely via PAYG fees (without relying on inflationary emissions which will run out), and does so with the elegance of both sustainability and the natural ties to L1 growth synergistically fueling main chain security. To anyone who answers “we’ll worry about that later when emissions run out”, another way of asking this question is: “When your validator set has declined by 50%, your total stake weight securing your network has also declined, and your token value has declined, and all three are still trending down long-term, how will you turn that around and improve your network?” The reason this phrasing is powerful is because is a description of the current situation with the Avalanche main chain. In fact, many well known, long-time validator operators and community members have publicly and vocally (and reluctantly) capitulated and exited the validation market once their stake unlocked, having not returned. The metrics show this is in fact the trend. Our emissions-driven validation incentives, which have steadily declined along with $AVAX price and utility, are detached from L1 growth and actually incentivizes stagnation or shrinkage of the validator set. They simply can not financially justify locking a declining asset, for low yield, etc. This dynamic is a death spiral that I am attempting to reverse with this (and future) ACPs, by introducing a natural market balancing mechanism. If someone answers “L1s will one day voluntarily incentivize main-chain validators with emissions”, I would say that the L1 growth linked to main chain security dynamic broke with the decoupling of the L1 validator sets in Avalanche9000. Validation metrics show this strongly as well. It’s all fine and well that L1s will likely contribute more directly to the validation market. I accept this, and in fact a big part of this proposal (outside of our scope of just discussing (1) here) is centered around this aspect, and the arising niche validation needs, etc.. However, they are still free to incentivize validators that are disjoint from the main chain, and without other aspects of this proposal, there is also arguably a "tragedy of the commons” or externalities dynamic, such that main chain security benefits all L1s, but no single L1 will foot the bill of it for others. This proposal addresses that by proportionally linking every L1 to incentivized main chain validation. I’ll conclude this main discussion with the first example with numbers, to illustrate what I’ve referred to above as the “market-driven, L1-aligned, self-balancing” effects, and its impact on $AVAX valuation: Example 1: Example 2: Validator rewards are distributed in proportion to stake weight of the validator (up to a delegation/stake cap). The more $AVAX staked with your validator vs competing validators means a bigger piece of that rewards pool pie comes your way. These rewards go to the stakers, minus a fee you can charge them as the validator operator. So $AVAX holders are now already incentivized to stake by delegating to a validator (or do so indirectly via liquid staking rewards), because the staking rewards will increase to a bit below the validator yields (23% in this example). Validators are themselves incentivized to attract stakers to further increase awards directed to their validator, and compete for stakers by keeping their fees low for example. (None of this dynamic is new.) So lets say I run a validator and drop my fee charged to delegating stakers to say 10% of the yield they would earn (so an additional 0.1 of 23% = 2.3% for me.) And this return does not even account for the appreciating price of AVAX as its demand increases to seekers of this yield. So now, much like a dividend plays a role (among other factors) in setting a fundamental floor value with stocks, its now very profitable to buy $AVAX at current prices to earn 20-25% yield (staker and validator, respectively). And because the stake and fees are both denominated in $AVAX, this yield doesn’t decrease as $AVAX price increases. In fact, in USD terms, it increases their $USD-denominated ROI (based on price when locked/invested). Example 3: However, none of this even accounts for greater order flywheel effects. For example, since AVAX denominated fee yields are now being seen through the market’s lens of an in-demand, vs. depreciating asset, the market pricing of “equilibrium” yield will take this improved asset valuation growth / floor support dynamic into account by assessing lower depreciation risk of capital, meaning that an even lower yield would be deemed attractive, further driving AVAX demand / staking participation, and lowering perceived risk of longer term locking (again adding price dynamics by reducing available supply). Yes, an offset to the benefits here (if all else like emissions are held constant, which they wouldn't need to be) is that the PAYG fees are not burned, so we lose a small amount of net deflation. However, note that the c-chain has doubled in terms of validators and/or stake weight. One flywheel effect here is that (Alternatively, since emissions is governance-controlled, this yield can also come down part of the way, from 25% to 20% by deciding to lower inflationary emissions from say 7% to 2%. Which in turn increasing value of $AVAX. Even here, we still need significant additional buy pressure to go from 20% to 10% yield at the assumed equilibrium.) Example 4: Note that if validators leave the pool, those that remain increase their proportion of awards incentivizing them naturally to stay at times when some may be lured by other opportunities. Again, the market mechanism keeps validators incentivized to not leave, with a natural validator balance being proportional now to L1/ecosystem size.… that's a rising floor as L1s continue growing. If core AVAX-locking validators and stake drops out of balance with the L1 ecosystem activity, incentives to come back up to balance immediately take effect naturally.. This gives L1s an additional assurance of continued stable security. Note: We do need to fill in more numbers and extend from the examples above into more complete scenarios. I’m also responding off-the-cuff and have not double checked the numbers above, much less barely proof-read them. I am erring on the side of quick responsiveness to the first part of the comment, for the sake of conversation flow/momentum and for illustration purposes. More rigor should be applied to these examples with time. A few final thoughts: Here’s the beauty of all this: Avalanche, by our architecture, is in a rather unique position to implement this vertically-scaling fly-wheel. Solana can’t. BTC cant. ETH doesn’t have the L1 PAYG . Note that they are focusing heavily on native interop so they could copy this proposal in the future, but they are BEHIND us in regards to having an architecture flexible enough to capture these fees from growth as cleanly/naturally. They’re still bound by fees for roll-up transactions, etc. And many of their L2s can simply grow their value very independently. By the arguments and examples above, we can see that, while AVAX holders (non-stakers) may at first glance appear to not benefit from (1), it is far from the case: The dynamic increases staking/validation yield (all else equal), which 1.) increase validation demand, which 2.) Staking demand, which 3.) increases $AVAX demand and locks more supply, which does benefit all, including simple holders. Finally, 4) With the now-direct linkage of yield to L1-growth, a forward-looking component of $AVAX price comes into play, much like PEG-like dynamic in equities accounts for future revenue growth linked to clearly measurable facets. So AVAX price can now account for projected L1 growth directly, and growth gets priced in now. This is good for AVAX holders, even those not staking. We note though that arguably there should be some comparative dis-incentivization to idly hold (vs stake) AVAX. We, as a collective, WANT the rewards to go to those who contribute/stake their assets to the network’s security. So by concentrating the PAYG toward rewarding validation/staking, it encourages much needed staking / validation and productive asset use. Finally, a quick note on some potential nuances of (1): The second point of nuance related to (1) is that here we focused on only the PAYG fees being used for rewards. Fees for Interchain Chain Messaging (ICM, and related interop capabilities) are arguably also strongly tied to the value main chain validators provide and could be used to link their incentivization (and hence AVAX value dynamics) to L1 growth. So a variant of (1) that we can discuss further is to include ICM (and related interop / usage) fees in what is directed to stakers/validators (vs. burned). It is arguably as elegant of a mechanism. Alternatively, they can remain burned. An argument for the later may be that we still get a burn dynamic that is also linked to ecosystem growth, while the PAYG fee rewards are “sufficeint enough” for market balance. As we start to run more numbers modelling this proposal, we may see, for example, that PAYG fees alone are not sufficient to move the needle in proportion to expected growth and existing stake weight. That’s when discussion of the ICM fees, as well as potential compounding effects with other parts of this proposal, including (3), and even other ACPs can be considered holistically. The PAYG fee proposal is just a start, but its importance is in introducing the validation market-linkage dynamic. I just wanted to raise this point of nuance, but will leave its refinement to beyond scope of this response. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
I've been working on a much more comprehensive set of ACPs around AVAX tokenomics (and specifically, its alignment with L1 ecosystem) which I plan to introduce when ready at MVP level, but separately want to propose a few related ideas to improve the alignment of the $AVAX token value:
Note: Current validator count (and weight) has fallen from a peak of over 1,860 to under 800 (weight from over 230M AVAX to apx. 150M AVAX) recently, and has been trending down, which is a concern for security and ecosystem health. By directing the growing fees from L1 adoption to reward the main validator set, we create a market/natural incentive mechanism for validators (and stakers) to participate in the health of the network.
A more nuanced version of this would allow the allocation of L1 fees to be split (and governance configurable, subject to hysteresis dynamics) between burn and direct validator incentivization. However, this may not be necessary/beneficial, since total emissions is already governable, and L1 fees can be an input to that governance decision.
Change from the current statically-determined fee mechanism (e.g., 1.33 AVAX, etc.) to instead be set (again, subject to some hysteresis meta-configuration dynamics) dynamically as a percentage of the total validator stake weight on the main chain. This now directly aligns the L1s with the quality of security (a mutual benefit/goal) they are receiving by participating in the Avalanche ecosystem, via simple market dynamics. For example, current total stake weight of C-Chain is apx. 150M AVAX. If we say that the baseline L1 fees are to be 0.000001% of total C-Chain Stake weight, this results in a 1.5 AVAX baseline fee by current metrics. If main chain security/quality decreases because we lose stake weight, then L1s pay less for the lower quality product. But as C-Chain stake weight increases, they pay more for the better quality network. (We would also keep, in tact, the current mechanism to scale fees formulaically from this baseline, based on L1 scale thresholds.) This mechanism works in conjunction with (1) above (fees going to validators), so as to keep rewards proportional to overall main chain validation growth to avoid reward dilution.
If we want to be truly institutional-grade, our main chain should aspire to this growth. The current decline in main-chain validation and stake weight needs to be reversed.
Note that more nuanced discussion can occur around other variables/dynamics that could enhance token/value alignment with L1 fee structure (and I have draft ACPs on this topic, not yet published), but this above dynamic can be introduced as a component independently and separately.
A final note: There are variants of the above dynamics, as well as a lot more detail (much of which I am drafting separately), but the intent of the above is to provide a minimum-viable introduction appropriate for this "ideas" phase of the ACP process, erring on the side of brevity to facilitate community discussion.
Beta Was this translation helpful? Give feedback.
All reactions