Inflation Focused LIP Discussion Thread

[vires-in-numeris]

From a perception and token utilization (e.g. in DeFi) standpoint, limiting the inflation is certainly desired. More DeFi integrations might also help with the Livepeer mindshare in crypto - something that we’re really lacking.

On the other hand, lowering the inflation could have various consequences like e.g.:

• Less participation
• Fewer Orchs (less inflation income, so no longer profitable)
• General raise of the Orch cuts (IMO not a bad outcome…)
• Less treasury budget

I guess most people also expect a higher higher/more stable LPT price if we lower the inflation. But I doubt that we’d see an immediate or even short term impact.

I haven’t formed a strong opinion yet and I’ll just throw another approach into the discussion that came to mind:

• Count the treasury towards the participation rate (makes sense since we use the treasury funds for participations in the Livepeer ecosystem)
• Increase the treasury cap and funding rate (which has a similar same effect as reducing inflation: Less rewards for Orchs and stakers and therefore less LPT on the open market)

If we’d e.g. increase the funding rate to 20% and count the treasury towards the participation rate, we could reach the 50% sometime in early July (assuming a 20% unbonding rate) according to my rough estimations. And after that, the inflation would start to go down again.

The advantage is that we can somewhat control the LPT issuance without making any hasty or arbitrary changes to the participation target and/or inflation rate. And of course that we’ll have a large treasury in control of the LPT holders that can be used for various initiatives to strengthen Livepeer’s position. There’s also the option to burn a portion of the treasury to reduce the LPT supply.

[dob]

Lots of good discussion in this thread. In this post I’ll briefly respond to some of the recent points, and then in a subsequent thread I’ll outline a potential step forward.

• Counting the treasury towards the participation rate and increasing its contribution rate: Definitely an interesting idea. But I don’t think the treasury is meant to be a sink that just accrues token to burn them or keep them off market. We’re trying to get more effective at allocating token out of the treasury to SPEs, and therefore it won’t have a positive effect on participation, and will hit the market. The idea of burning funds from the treasury is interesting if the community wants that, because all it requires is a proposal on the treasury to send LPT to a burn address if passed. (Though we may want to think about the perceived impact of this because it won’t technically reduce the reported supply…which is what the market seems to look at if they don’t understand the nuance of burn addresses).

• There were a bunch of comments about better L2 access to the token helping to increase participation rate. Binance now supports L2 deposits/withdrawls of LPT. There’s definitely opportunity to market this to increase participation. But it’s unlikely to have an overnight short term impact that gets inflation trending downwards.

• It does seem like there’s a pretty wide held view that reducing or capping inflation brings benefits.

• There’s a thread of discussion around the project and model being failed because it doesn’t generate enough ETH fees yet. Anyone is welcome to this view. But tech startup history is littered with countless examples of 8, 10, 12+ year “overnight successes”, where the groundwork was laid for many years to littler awareness or growth, before the timing, or the market, or the product was right for it to catch fire overnight. We have an incredibly solid foundation, and Livepeer’s big bets on realtime AI, agents, Daydream, etc are all attempts at dramatically growing fees. No one has to stay along for the journey, but there are a lot of people here in the community working hard on demand generation!

• There’s a thread of thinking around O’s being able to cover their costs or not as inflation comes down. O’s have multiple levers available to them, including reward cut, fee cut, their own hardware and utility costs, etc. None of the proposals suggest inflation coming down over night, and O’s can certainly adjust their parameters to ensure they’re meeting their costs at varying inflation rates. Of course fees need to ultimately increase to make the network sustainable for a large number of O’s…but a large number of O’s are only needed if there’s a large amount of work and hence fees.

• Lastly, a reminder that inflation isn’t designed just to cover O cost to break even. In a world of token distribution mechanics that are rife with weaknesses and risk (crowdsales, ICOs, airdrops, etc), LPT inflation which distributes token over many years to those who are showing up to help the network, has worked quite well as a long time horizon distribution mechanic. It is partly routed by delegation and reward cut, and it has shown to be a mechanism that has helped build up LPT stakeholdership amongst those who are powering and building the network. This is resulting in many new committed builders and organized demand gen shots on goal, and should continue until the network fees themselves are enough incentive to build token stake to compete for fees.

[dob]

I’d like to move forward with drafting an LIP based of the framework that @stronk proposed: Introducing a inflationCeiling and inflationFloor value. Since we’re likely above the ceiling value, inflation would tick down by inflationChange each round until its within the proposed range, at which point our standard participation target rules would apply.

This will give the community two more parameters to modify that play into the economics. And it would avoid arbirtraily straying from our 50% participation target. While I proposed a couple sample values that could be considered as starting points, I think the community’s job between now and the final proposal going live on chain in 14-21 days, would be to suggest param values for the:

• inflationFloor
• inflationCeiling
• inflationChange

If people really wanted to see the current incentives play out, we could set the ceiling to be higher than the current rate. This would have no immediate impact to reduce inflation, but it would still cap it below what people would consider runaway hyperinflation. Ultimately in time we’d hit 50% target most likely and then inflation would begin to reduce.

If they want to see it set to something more in line with DePIN or DeFi norms, we could set it near the 10% annual max level suggested. And of course, when these params exist, they can always be moved via anyone’s parameter change LIP proposal and vote.

[Joel]

My feeling would be to go with the first option…ie setting the ceiling to be higher than the current rate and preventing runaway inflation. This changes things in a positive way, while not causing any major disruptions. Ultimately, it doesn’t feel like there is a high degree of confidence in any of the proposed changes, and in light of this, might be best to stick with the less intrusive adjustment.

[drieddate_sidestream]

Hi everyone!

We’re Sidestream, and for the past several years we’ve been serving as one of the core protocol contributors at MakerDAO/Sky with the focus on governance payload security and protocol upgrades.

Recently, we’ve been discussing the opportunity to play a similar role in the Livepeer ecosystem. To showcase our ability to deliver meaningful value, we selected the currently pending Minter upgrade PR, which introduces inflation bounds, and enhanced it with the goal of bringing this update to production.

More specifically, we’re adding multiple unit and integration tests, as well as a new function to seamlessly migrate the Minter state while ensuring that no rewards are lost. You can find the full description of what was done and review our work in the PR itself.

We’ll also introduce ourselves in # protocol within Discord, so feel free to get in touch over there.

[Authority_Null]

I’ve come back to this thread several times and have tried to piece together my thoughts but am really struggling to put together something constructive.

The ultimate goal of the network (I thought) was to increase participation. It seems like the idea suggests the potential negative perception of high inflation may be contributing to less participation, but I’m not sure this is the case. As @dob mentioned participation decreased after the move to Arbitrum, likely because of increased bridging times mixed with unbonding times, bridging in general, less interest in yield farming globally, and an overall convoluted process to get involved.

To me, and I mentioned this during a watercooler a while back, the points above are the real issue. The proposal is treating symptoms, not the problem, which works for Tylenol, but I don’t know how it will work here. My speculation is that participation won’t increase based on these changes, but orchestrator cuts will go up, while fees stay the same, unless SPE’s start really driving traffic.

Again, I’m just trying to wrap my head around this, but I’m looking at it from a very high level and may be missing out on the technical scope and improvements that can come from it.

[Karolak]

Equilibrium has yet to be found. If current rewards are not enough to outweigh the benefits of staying liquid, then lowering rewards further will not attract more delegators. Setting a strict ceiling—such as 20%, which is below both current inflation and equilibrium—could put the network at risk, preventing it from ever attracting new delegators. The main driver—higher rewards that compensate for illiquidity—would be removed.

This is why I support the current system and oppose introducing rigid limits. Let supply and demand determine equilibrium. If fees are flowing, staking will become more attractive, naturally reducing inflation. The system is working as intended: when network usage is low, inflation increases diluting unstaked capital to subsidize orchs; as usage rises, inflation decreases eventually increasing LPT price as demand for staking to get % of the fees will rise.

The only parameter worth adjusting, in my view, is the inflation rate, as it could help reach equilibrium more quickly. We’ve already eliminated the seven-day bridging time by integrating Binance. In reality, a single CEX is enough to sustain DeFi liquidity. Arbitrageurs operate swiftly, automatically bridging LPT from CEX to L2 to close price gaps. Combine it with cowswap limit orders, you can enter/exit pretty quick!

I recommend keeping an eye on Binance’s Arbitrum LPT hot wallet. Demand for LPT consistently outpaces supply, signaling that we may soon reach 50%.

Binance Arbitrum LPT

[vires-in-numeris]

I agree. I don’t see any immediate issues with the current inflation (besides maybe DeFi usage, but realistically we’d also need other changes to fix that).
And since we’re now only 1.6 percentage points away from 50% participation, introducing changes with unknown consequences doesn’t seem to be right approach at this moment.

[dob]

Glad to see the participation rate up over 48%. Hopefully we’ll hit our 50% target shortly.

In the meantime, I’ve authored and submitted a draft LIP that:

• Introduces the inflation ceiling and floor.
• Suggests they be set to about 30% and 2% annually.
• Doubles the inflationChange from 500 to 1000 to speed up the rate of change in either direction.

@drieddate_sidestream has kicked off the security work of testing the proposed change and simulating its deployment.

The motivation behind moving this proposal through the process is that it will introduce controls that can be used in the future to set the cap and ceiling via governance should the community want to. It will not effect the current inflation trajectory or incentives. But it will prevent “runaway inflation” beyond what the community might deem a tolerable level. And speeding up the inflation change will allow us to find equilibrium in either direction quicker (theoretically).

I think these benefits are nice, but given the above comments about the inflation working as intended, and us being near the target, I understand that it is not necessarily critical to pass this if the community feels that the risk of the update or introduction of these parameters bring other drawbacks. We’ll see through discussion and ultimately a vote, how the community feels.

View the draft LIP PR here.

[obodur]

I think increase or decrease even faster every round instead of doubling it but no ceiling would be better because otherwise if 50% participation is not reached until inflation arrived at the ceiling, that might discourage (or at least not encourage) people to delegate more and that could be even worse for the network overall.
In short, let’s make the change amount even higher (maybe 4x the current amount) and scratch the ceiling too.

[dob]

Livepeer Inflation Path Forward

Earlier this year an active discussion kicked off regarding the high current rate of inflation on the LPT token that powers the Livepeer network. With rates approaching 25% annually, and continuing to increase, stakeholders were concerned about issues such as the possibility of runaway inflation, overcompensating node operators, and potential participants’ negative perception of the tokenomics.

The reason for the continuing increase of inflation rate was that Livepeer inflation adjusts automatically based upon whether greater than or less than 50% of all active LPT is staked and participating. The network had been hovering between 40-44% for a number of months, and it had been over a year since the 50% target had been surpassed, leading to constantly rising inflation. In the iterations of the community discussion that followed, a couple options were suggested, and ultimately, a candidate solution was devised.

It was proposed in LIP-100 that:

• An inflationCeiling and an inflationFloor be set to limit the bounds of inflation. These values would be set via community governance.

• The rate of inflationChange would be doubled, to hasten the rate that inflation adjusted itself within these bounds.

The effect of this would be that the inflation rate would essentially be bounded within limits the community and market deemed acceptable, preventing the risk of runaway inflation, and trending itself towards this rate without any immediate jolting change to the economics for current node operators or delegators.

This proposal was written out, implemented, as well as tested and simulated successfully by @drieddate_sidestream. It stands waiting, and ready to be proposed for community vote. However, prior to moving to vote, it is worth noting two other key facts that emerged regarding this proposal.

1. The participation rate started to increase towards the 50% target. The high inflation rate was doing its job, allocating more and more stake to those who were participating. The 50% target has actually been achieved a couple times, and inflation was starting to tick down on its own.

2. The proposed upgrade is not without risk. It requires migrating all the on chain staked LPT and ETH in the protocol ($100’s of millions of dollars worth of value) from one smart contract to a new one. While scriptable, testable, and done before, this obviously can’t be taken lightheartedly, and isn’t without risk of bugs or exploits. There are a number of ecosystem actors like exchanges, explorers, and staking tools that read values from the old contract such as supply, staked amount, etc that would need to be upgraded, and there’s always risk they drop support or appear broken for LPT. With recent positive momentum around LPT token integrations, it would be a shame to break them or risk unlisting activity due to lack of their ability to make timely upgrades on their end to reflect a new contract.

While still probably a good idea to consider introducing these parameters for the future, it has felt less urgent to the community to undertake this slightly risky update, being that inflation was already coming down on its own. The upgrade stands waiting and ready to be proposed on chain and voted on, should inflation start to rise again, and should the community wish to see it come to a decision.

The next steps

Just because inflation started to decrease however, doesn’t mean that all the underlying issues are solved. After a recent spike in token price, significant unstaking occurred, and it is likely inflation will continue to rise again. People could argue the rate of inflation will still remain high for a while, even if it begins decreasing, leading to negative market perception and a lack of interest in participating in the network. The community is left with at least these three options:

1. Vote on the above proposal. If it passes, undertake the slightly risky on chain protocol upgrade. This only requires a community member progressing the last step in the LIP process to make this available for on chain proposal, and proposing on chain in the explorer.

2. One alternative suggestion here is doubling the rate of inflationChange and dropping the participationTarget to something lower than its 50% value. This simple parameter change proposal would be technically risk free relative to the more complex code upgrade, and it would mean that inflation will tick down from its currently high level nearing 30% annually, to below 10% annually in approximately a year, if the participation target continues to be exceeded. This is a small part of the existing proposal, which can always be voted on and enacted in full at any future point. However it does come with a seemingly arbitrary change to the target.

3. Do nothing, and let the participation rate target of 50% work its magic, and the inflation rate slowly adjust over time based upon the current parameters. If the participation rate sinks well below 50% and people are concerned, they can always proceed with voting on the proposal then.

It is also worth noting that the new Foundation and its advisory boards are actively working on creating a new protocol security and development SPE. One action that this SPE will likely take is to future proof the protocol, make the Minter contract proxy upgradable, meaning that future upgrades of this type won’t require the onchain migration of funds that this risky upgrade would require. It may be prudent to take only a safer, short term action, in advance of this, and wait for the security work to proceed before dealing with the pain of the onchain funds migration once and for all.

While there’s a lot that can be addressed in the protocol overall, including the all important necessity of increasing the amount of utility based fees on the network relative to inflationary rewards, the above is aimed at giving network participants the confidence that the protocol is operating at a more sustainable level of rewards and inflation going forward, likely bounded within a defined and understandable range.

My suggestion would be considering the drop in participationTarget, perhaps to 40%, and double the inflationChange. As the protocol SPE comes into form, consider the security work to make the Minter contract easily upgradable, and then introduce the ceiling and floor bounds to lower values than we’re currently seeing. Ultimately it’s up to the LPT community to decide and vote.

[cult_leader_en]

Option 2 here seems to be the easier short term path. I am a fan of dropping participationTarget to 45% and 2X the inflationchange !

[Karolak]

@dob

inflationChange

Option 4. Separate voting for inflationChange and participationTarget. I can imagine one may be for inflationChange parameter change but not for participation target and vice versa

[obodur]

Increase inflationChange.
Do not touch participationTarget!

[Pon]

I would say change participicationTarget to 40 % and change inflationChange to x4.