Token vote locking has been all the rage for quite some time now. Pioneered by Curve, protocols like Ribbon, Yearn, Hundred Finance and many more adopted the model. While there are definitely advantages, the downsides are rarely talked about. Instead, it seems like every week a new project team announces their switch to a ve token design, which is seen as a cure-all for everything from low TVL to bad price performance.

Let me play d̶e̶v̶i̶l̶’̶s̶ bear’s advocate: What’s bad about ve tokens and what could potential alternatives look like?

Why protocols adopt the ve token model

How ve tokens work

ve stands for vote escrowed

• Escrowed: Users choose how long they want to lock-up their tokens (e.g. CRV). It’s impossible to unstake early. Tokens unlock linearly over the selected period. Users may perpetually lock their tokens to max their voting rights and rewards.
• Vote: Proportionally to their lockup period, users receive an amount of ve tokens (e.g. veCRV) in return. ve tokens are used for governance, meaning that community members that lock their tokens for longer have more weight. The same often applies to rewards. For example, locking CRV for the maximum of 4 years grants users a 2.5x boost on their LP rewards.

To summarize, stakers lock their tokens for a desired period and receive voting (and reward) rights, usually proportionally.

What benefits teams see

Locking shows the long-term commitment of stakers. They are willing to turn their stake illiquid (-> opportunity costs) in return for voting rights and rewards. This should improve governance as committed members have more weight in decision-making. These stakers should a) be interested in the long-term health of the protocol (incentive alignment) and b) be the ones that are willing to make informed, time-consuming decisions (self-selection).

These community members are more valuable to a protocol than others and therefore also receive more rewards.

Downsides of ve tokens

Lower-than-expected incentive alignment

When prices go up, the world seems ok. But once goblin town calls, the game theory turns dark. It’s rational for stakers to maximize their personal gain (e.g. receiving the biggest bribes) rather than make decisions supporting the long-term health of the protocol. This behavior especially takes place when sentiment turns bearish. Who knows what the tokens will be worth in 4 years. Better to take with me as much as I can now. I already fully depreciated my principal mentally. The others can take a long-term time horizon for their votes.

Convex-style protocols on top of ve tokens take the “incentive alignment via locking” ad absurdum. Another protocol soaks up vote-locked tokens like a hungry Kirby. This takes supply out of circulation and reduces sell pressure. On the other hand, the locked tokens are governed by holders of the meta governance token with usually a shorter lockup period. Meaning, the long-term alignment is gone. You could argue that the meta token holders don’t want the underlying token to fail as it makes them revenue. But as we know, every individual looks after themselves and who knows what will happen in crypto in just a few months, let alone years.

So, ve tokens might not help as much with long-term alignment as first thought. 

Building a s̶t̶r̶o̶n̶g̶ community*(*of bagholders)

Once the trust in the ve token model erodes further as people realize that there’s limited long-term alignment, they are looking to sell. Problem for the most committed community members: They can’t - their stake is locked for another 69 months. The stakers that followed the golden path of ve tokens, locking their tokens for the max period, are the ones that are rekt the most. Devastating for them, innit. And not helpful for community building.

The lockup works in both directions. While it helps to reduce sell pressure by taking tokens out of circulation, it also makes the token more unattractive to buy. If I have to lock the token for 4 years to make the most out of it am I really gonna buy it? The opportunity costs in a dynamic environment like crypto are immense. To be fair, liquid wrappers tackle this issue by tokenizing locked stakes. But as described above, this reduces long-term alignment. Also there’s the issue of the tokenized position depegging from the underlying ve token. The “tokenized ve token / ve token” pair is the main pool and exit liquidity is often low.

Example
cvxCRV holders exiting means cvxCRV gets dumped for CRV, then CRV gets sold for ETH/DAI/… cvxCRV holders are both affected by the CRV price decrease as well as the cvxCRV depeg.

Moreover, the meta governance token might look more attractive, stealing demand from the ve token.

Especially for high-emission tokens, locked stakers will feel like they are constantly getting dumped on. This influences their behavior. They become more likely to seek short-term profits. Others observe this and start acting in the same way. Trust erodes more. A vicious cycle.

Governance halt

The issue extends to the protocol level. Demotivated stakers want to leave but can’t. They are locked. They still hold ve tokens that act as governance votes. The chance that they are interested in any governance decisions not involving short-term profit is low.

Their voting power only decreases slowly over the lockup period. At the same time, external capital will hesitate to buy and lock tokens for long time frames in such a situation.

This makes governance difficult in times where it is needed most and may even completely halt it.

Meta governance tokens lead to a cartelization of ve voting rights but as long as the meta governance power is distributed to token holders, it doesn’t add additional risks aside from the issues of liquid wrappers mentioned before.

Alternative tokenomics

To have token holders more long-term aligned, protocols can use different alternative tokenomic mechanics. However, in the current growth stage of crypto with new narratives and projects appearing almost every day, it may be difficult to retain a majority of community members and stakers over time. Tokenomics can help, together with other measures of community building.

Protocols need to be aware that locking in stakers might not be as valuable as first thought. You cannot buy loyalty. Dissatisfied stakers make bad governance decisions.

Therefore, I argue for soft lock-ins. These add opportunity costs for stakers with the goal of reducing short-term fluctuations and inspire people to engage in long-term thinking, but allow them to leave anytime. Only happy community members are valuable community members. 

Let’s look at two token mechanics and their trade-offs, shall we?

Time-based tokenomics

GMX, Platypus, Prism Protocol and Kalao all make use of time-based tokenomics, although with slight variations.

Stakers accumulate boosts over time which award them more rewards and/or governance rights. They can unstake anytime but then lose their boost, which constitutes opportunity costs. Opportunity costs come in the form of lost future revenue and voting weight. This can help reduce short-term volatility. For example, stakers are less likely to unstake during a price drop they expect to be temporary. Same applies to weeks with lower protocol revenues. As long as the medium- to long-term outlook is positive, users will remain staked.

On the other hand, stakers are flexible and can leave when their longer-term evaluation changes. They can react to new information, such as problems in protocol adoption, new competitors or new demand in the market. As a single (non-whale) voter has limited influence, it is fair to give them rights to exit. For the protocol, this comes with the advantage of faster recovery of governance processes.

Compared to ve tokens, which are considering a staker’s commitment into the future in the form of lockups (forward-looking), time-based economics are backward-looking (how long a user is already staking). Past behavior is a predictor of future behavior, but of course not perfect. Nevertheless, the longer a user has been staking (chose to not unstake under different conditions), the more likely that they will keep staking in the future. The opportunity costs are a further incentive to remain staked.

Variables

For every token model there are trade-offs. We can decide the balance by setting certain variables.

With time-based boosts we have to balance the advantages early stakers have over new buyers. The bigger the advantages, the higher the opportunity costs of unstaking. This reduces sell pressure but also makes buying late less attractive (lower buy pressure). Additionally, it leads to a higher concentration of voting power (when users receive the native token as a reward) and revenues to early stakers. Governance recovery takes longer as long-time stakers have higher weights. On the other hand, a long staking history indicates that a user cares about a protocol and also increases the chances they are well-informed, both of which are raising decision-making quality.

The following factors determine the early staker advantage

• Max cap
• Time required to reach the max cap

The higher the max cap and time required, the higher the advantage of being early.

There’s also the unstaking penalty, for which two models exist. Platypus’ burns all accrued vePTP when a user unstakes any amount of PTP (full penalty). This increases opportunity costs. GMX burns multiplier points proportional to the share of tokens a user unstakes (partial penalty).

Protocols can also introduce an unstaking period to further disincentivize short-term trading. Prism Protocol has a 21 day unstaking period for xPrism.

While Platypus and Prism employ separate reward pools for time-based boosts (base pool & boost pool), GMX and Kalao have one pool. Separate pools allow a team to ensure that non-boosting users receive a certain minimum APR. With a one pool structure, boosted users crowd out non-boosted users’ rewards. Token boost factors would need to be changed, while the two pool model has another lever via adjusting the reward weights between pools.

Reward vesting

The model of esGMX and esKLO is particularly interesting to me. Stakers don’t earn the native token but rather an escrowed token. They can decide to

1. keep the es token (the autocompoundooooor). The es token gets automatically staked and earns rewards just like the normal token
2. vest the token linearly over a period of time (the vestoooor). During this time, stakers won’t receive rewards on their es tokens. Vesting requires staking the average amount of tokens used to acquire the es tokens

This brings several advantages

• Opportunity costs: Stakers have to decide between vesting their es tokens or earning protocol revenue with them. Therefore, es tokens are especially relevant for revenue generating protocols.
• Effective supply reduction: Because of the staking requirement for vesting, not all es tokens will be converted into the native token.
• Decrease immediate sell pressure: es tokens cannot directly be sold. While it may seem like a “kicking the can down the road”-type solution to emissions at first, it can be helpful in the grand scheme of things. Token rewards are used to bootstrap supply and/or demand but can only be sold later after the protocol (hopefully) has established itself, meaning price is likely already higher and there should be more demand to soak up the sell pressure.
• Revenue share to long-term aligned stakers: More rewards get distributed to stakers that do not vest their es tokens.

On the flipside, es token holders still earn a part of revenues even after they already unstaked their principal. It’s fair to them as they earned the es tokens fair and square and still own them. But as they are no longer interested in the protocol, it feels like lost revenue.

Conclusion

Finding long-term committed stakers and contributors is difficult. While a hard lock-in secures people’s commitment on the capital side, it may not be the best way to incentivize behavioral commitment. Protocols should do more experimentation, especially with soft lock-ins, like time-based tokenomics and es tokens rewards. Of course as always, these token mechanics can be built in a way to maximize long-term sustainability or to supercharge the ponzi. As a builder or researcher, it’s valuable to look at the design space and trade-offs to come to your own conclusions. Also keep in mind that tokenomics are just one piece of the puzzle to long-term community retention.

👉 If you enjoy posts like these, short-form twitter threads and occasional shitposts, follow me @0xKepler