Optimal Token Vesting Models

Motivation

We’ve been spending a lot of time thinking about token vesting recently. 

At its core, token vesting is all about incentive alignment. How do you get all the key players—team members, early investors, and future investors—to pull in the same direction? This question is well understood in traditional equity structures but has not been as well defined within crypto.

The industry has experimented with a number of models, from simple linear unlocks  to complex milestone-driven models. Yet, a universally accepted "best practice" remains elusive.

We looked at past data to shine some light on this topic. Given our focus as a fund, this analysis focused on infrastructure protocols - it would be interesting to compare the data across a wider range of categories in the future.

Model Variations

There are 2 basic designs which make up most of the vesting schedules that have been implemented. Typically, there is some combination of the following:

Linear Vesting: A fixed number of tokens are released at regular intervals (e.g., monthly) over the vesting period until the total allocation is fully vested. This is a straightforward model that provides a predictable release of tokens.

Cliff Vesting: This involves a waiting period (the "cliff") before any tokens are released. Once the cliff period ends, a lump sum of tokens becomes available, and after that, tokens may be released linearly or in another specified manner.

Prior to the 2021 bull market, vesting schedules tended to be more front loaded - investor tokens tended to be mostly unlocked within the first year. However the recent trend has been towards delayed cliffs followed by additional linear vesting. The result is fewer tokens in circulation which can lead to inflated fully diluted market caps - e.g. Aptos, Sui, Worldcoin, and others.

We wanted to understand the pros and cons of these different models and focused on 2 key questions:

• How does the market react to cliffs? Is there any clear benefit to having a 1 year cliff versus a 3 month cliff?
• What is the relationship between vesting schedule and the price performance post TGE? In other words, is there a noticeable difference in tokens which have an aggressive unlock schedule (i.e. investors fully unlocked within 1 year) versus a conservative unlock schedule (i.e investors have a 1 year cliff and 2 year linear after that)?

How Does the Market Treat Cliffs?

We examined 15 protocols which incorporated cliffs in their vesting schedules. For each protocol, we looked at the price data 30 days prior to the cliff and 30 days after.

The data pretty much lined up with what would be expected. Token cliffs negatively affect price. Some key takeaways:

• Nearly every token performed poorly in both directions, but they performed significantly worse leading up to the cliff, indicating that the market prices cliffs in, in advance.
• We looked at 3, 6, 8, and 12 month cliffs and they all followed similar trends. It appears that cliffs negatively impact price irrespective of how far out they take place.

Case Studies

Let's examine two protocols with different cliff lengths.

Flow (3 month cliff)

Before: -25% After: -18%

Optimism (12 month cliff)

Before: -33% After: -4%

As you can see, both FLOW and OP show quite similar price behavior leading up to their cliffs despite OP having a 12 month vest and FLOW having just a 3 month vest.

The above data would seem to suggest that based purely on price performance, there is no significant benefit to having a cliff be 12 months out rather than 3 months out. A logical next question would be why have a cliff at all? Why not just start vesting at token launch? That leads into our next observation.

Token Launch

We collected price performance data for 25 protocols.. For each protocol we assigned a category based on the aggressiveness of their token unlock schedule. They were labeled aggressive if fully unlocked within the first year or so, conservative if 3 or more years, and moderate if in between.

The charts below show the average performance for each cohort over time. In order to account for the underlying market, we took the performance relative to BTC+ETH during the same period. This means that any value above 1x outperformed BTC+ETH and any value below 1x underperformed. In the second chart, we excluded SOL, MATIC, and RUNE given how much they skewed the Aggressive group.

The data is not overwhelmingly conclusive, although it  suggests that the more conservative token schedules underperform compared to the more aggressive schedules - especially over the first 12-18 months. After two years, the token vesting model doesn’t appear to impact price at all (when excluding outliers). In fact, none of the groups outperformed BTC+ETH after the 2 year period. 

It’s worth pointing out that there may be some level of time bias in this dataset given that we only included tokens which launched 2+ years ago. As a result, this BTC+ETH outperformance may have just coincided with the previous alts cycle. However, as you can see from the below table, each group was fairly diverse in terms of launch date.

As mentioned earlier, it appears that more conservative vesting periods underperform relative to the other categories. This has interesting implications given that a purpose for more conservative vesting schedules is to reduce sell pressure on the token. It would seem that this is not how things play out in reality. A possible explanation for this is that tokens with a massive supply overhang are less appealing to investors. In other words, why would an investor want to buy at launch when they know the token will experience rampant inflation in the future?

However, this model seems to have been a bit of a trend with recent protocols - Worldcoin, Sui, etc. The benefit of this model is that the FDMC at launch is incredibly inflated. Worldcoin launched at a $20b+ FDMC and Sui at $15b - each has a 1 year cliff with multi year vesting. However, both of these coins have been down and to the right since launch. 

With that being said, both projects are still worth well into the billions. So perhaps this model works fine if the goal is to maximize the FDV at launch. Whether or not that should be the case is up for debate.

Conclusion

It's important to highlight that there are countless other variables which play into a token’s price performance. The token unlock schedule is just one. The goal of any vesting schedule should be to align the interests of the team, the investors, and the community. With all that being said, our observations led us to the following conclusions:

• Cliffs will always end up negatively affecting price. Why not get rid of them entirely?
• There appears to be no difference in long term performance based on when the initial cliff occurs
• More conservative unlock schedules may be detrimental to price action through the first couple years. But ultimately, the token vesting model does not appear to affect price in the long run
• If the goal is to optimize for token price to peak at launch, then conservative models may still make sense. If the goal is to optimize for growth and bring in new stakeholders, then a more aggressive model likely makes sense.