Introduction

In early DeFi projects, it was a common practice to reward early users through liquidity mining to facilitate the cold start.

This method contributes a certain amount of initial liquidity, but it cannot underpin the sound development of a protocol in the long run. Intuitively, users are still speculative and think keeping their money in their pockets is safe, so they mine, withdraw and sell. As a result, there will be an inevitable rapid decline in the protocol, which is well depicted by the death spiral of defidigital.

Besides, when the tokens of a protocol can only be used for governance, and the governance authority is quite limited in actual scenarios, users will also have insufficient motivation to keep holding them.

In summary, it is the problem of how tokens provide sufficient holding incentives, and how the interests of users can be kept consistent with that of the protocol for a relatively long period of time so that there is motivation that can contribute to the long-term growth of the protocol.

The emergence of ve (vote escrow) can well solve the above problem. After ve pioneered by Curve, many other protocols have incorporated ve into their own economic models, and have made their own iterations and innovations based on Curve’s ve.

This article will first explore the operation mode of ve, and take Curve as an example to analyze the pros and cons of ve. After that, we will combine the innovations of different protocols at different levels to explain the iteration of several core dimensions of ve. Finally, based on the above analysis, we will offer some suggestions for putting ve into practice.

Introduction of ve

1.1 How ve works

veModel’s core mechanism allows users to obtain veToken by locking the token. As a non-transferable and non-circulating governance token, it allows users to obtain more veTokens if they choose to have the token locked for a longer time(usually there is an upper limit on the lock-up time). According to their veToken weight, users can obtain the corresponding proportion of voting rights.

Users can use part of their voting rights to determine the ownership of the liquidity pool that can issue additional token rewards. This will exert a substantial impact on the user’s immediate benefits as well as enhance the user’s motivation to hold currency.

For the protocol, locking can effectively reduce circulation and selling pressure, thus making the currency price more stable. Meanwhile, after users lock their tokens, their vital interests are more consistent with that of the protocol during the lock-up period, which also helps them make the right governance choices. For those who are willing to lock their tokens for a long time, they can get more benefits and governance rights, which is also very fair morally.

Let’s learn more about the practical use of ve through the Curve model.

1.2 Curve: Pioneer of veModel

Trading fees are the revenue source for the Curve platform.

The platform charges trading fees during the process of swap (0.04%, collected in the token of the Output), deposit and withdrawal (0-0.02%). Among them, 50% of the trading fees are granted to veCRV holders (used to purchase 3CRV, and then distributed to the holders, who can receive it once a week), and 50% is given to LP providers (Base APY).

CRV token is native to Curve. This token is used to reward liquidity providers who stake LP tokens (liquidity provider tokens, namely, the token obtained by users who have contributed liquidity to the pool) into Gauge. Controlled by Curve DAO, Gauge controller is the core component of Curve. The Gauge controller records all parameters pertinent to voting governance, such as “the current slope of each user in each Gauge, and the power that the user can use and the time when the user’s voting lock ends”, etc.

The initial supply of CRV is 1.273 billion. Its total supply is 3 billion, and its initial daily increase is 2 million. These additional CRVs will be distributed proportionally to users who stake LP tokens to the Gauge based on the measurement results in the Gauge controller. These users own weight in the Gauge.

https://github.com/curvefi/curve-dao-contracts/blob/master/doc/README.md

When the staking users further lock the CRV, they can obtain veCRV (veCRV is non-transferable and can only be obtained by locking). The lock period that users can choose ranges from 1 week to 4 years. The longer the lock-up time, the higher the conversion ratio of veCRV. If a user locks 1 CRV for 4 years, he can get 1 veCRV; if 1 year, he can get 0.25 veCRV. During the lock-up period, the number of users’ veCRV will decline linearly as the remaining lock-up time decreases. If a user locks CRV multiple times, he must select the same expiration time.

veCRV holders own the governance rights of Curve. To encourage users to participate in governance (namely, encourage them to acquire more veCRV), they enjoy some of the rights:

1.transaction share

Users share the trading fees of the Curve pool equally with the liquidity provider with a ratio of 50:50. The fee is charged in tokens of output, for example, if users exchange USDT for USDC, they can get USDC.

What’s different is that the liquidity provider receives a direct fee reward (given in the same token of the liquidity provider, namely, Base vAPY), while the fee revenue of the veCRV holder will be converted into and distributed in 3CRV (3CRV is the LP token of three pools in Curve consisting of DAI, USDC and USDT.

The handling fee will be converted into USDC via Curve or Synthetix first, and then deposited in 3Pool. Later, 3CRV can be received when each veCRV holder account is updated.

Moreover, veCRV holders also have the opportunity to receive token airdrops of other projects that cooperate with the Curve platform.

2.Voting rights

In addition to regular governance voting, users can also vote to determine the attribution of CRV additional issuance. In other words, they can vote to guide the weight of additional rewards for each Gauge, which will affect the number of rewards that users can get.

Users can choose a lock-up time of t_l (t_l < t_max, t_max = 4 years), and their voting rights is w, the number of veCRV and proportional to the remaining lock-up time (t) and the number of CRV (a). For example, users A and B have the same CRV. User A locks them for 2 years, while B only locks them for 1 year, then initially, A will own double the voting rights relative to B.

Gauge will record the value of w when users operate (such as storing, withdrawing, etc.). This value remains the same until the user executes the next action. Since w declines over time, users can maintain their voting rights and rewards unless they lock more CRV and don’t operate during this time. If w decreases to 0, the user will be removed from voting.

3.InCRV bonus boost

Yield is the core indicator that investors care about. veCRV users can obtain a bonus boost. The more veCRV, the higher the reward coefficient. The maximum can be 2.5 times. It will further enhance users’ motivation to lock tokens.

The design of the reward multiplication mechanism is relatively complicated. Here is a detailed explanation:

1) The user’s balance in the Liquidity Gauge increases with the amount of veCRV owned by the user:

The aim to set a limit on the reward multiple of users participating in lockup is that it is necessary to prevent the veCRV whale from grabbing more benefits without limit even if the ve mechanism is designed to encourage users to lock and reward them for their loyalty and stickiness.

Taking the deposit of 100U stablecoin as an example, to achieve 2.5 times the income, 2840 veCRV is required, which is equivalent to 2840 CRV staked for 4 years. If calculated at CRV = 1U, it is equivalent to holding CRV with 28 times the value of LP Token.

Ordinary users will find it difficult to obtain so many CRVs in a short period of time, so there is often a limit to the reward multiplication. The reason why it was finally set at 2.5 times is probably because the team believed that a relative incentive balance could be achieved within this limit. Additionally, if a user provides liquidity to multiple pools, the reward multiplier may be different, which depends on the lock-up of LP in different pools.

The following figure summarizes the different rights and incentives for holding CRV and veCRV in the Curve system:

https://resources.curve.fi/crv-token/understanding-crv

Interpretation:

• Liq in pool ⸺> earns lending & trading fees: Provide liquidity (deposit) to earn handling fees

• Has veCRV⸺> vote on DAO proposals & vote on gauge weights & earn gov fees: hold veCRV: have governance authority, vote to determine gauge reward weight

• Liq in gauge⸺> earn CRV: stake LP token in gauge to earn CRV

• Liq in gauge & has veCRV⸺> boost: Stake LP token in gauge and still hold veCRV to get a bonus boost

Seen from the current operating results, this mechanism is quite successful. 45% of CRVs are voted and locked, with an average lockup period of 3.56 years.

https://dao.curve.fi/releaseschedule

Comparing the lockup period distribution, we can clearly see the long-term lock effect of ve.

https://members.delphidigital.io/reports/an-alternative-implementation-of-vetoken-economics

1.2.1 Advantages

We have mentioned the advantages of ve above. However, here is a summary of them:

1) After the lock-up, the liquidity decreased, which reduces the selling pressure and helps to stabilize the currency price.

2) Better governance possibilities: Governance rights are directly linked to yield distribution, and the higher the users’ motivation to hold coins, the higher the governance participation. Meanwhile, long-term staking users own greater governance weight, and they are also a group with the motivation to make better governance decisions. The time- and volume-based weightings reflected in governance weights currently appear to be quite reasonable.

3) The long-term interests of all parties are relatively coordinated: during the lock-up period, users are unable to transfer their tokens. Due to their interest correlation and rationality, they are more likely to ignore short-term and timely benefits and make better decisions that are consistent with the protocol’s long-term benefits.

The lockup mechanism also increases the short-term manipulation costs for giant whale holders. And once they choose to lock tokens to increase their voice, it is highly probable that they are prone to making rational votes that are in line with their own interests, and the possibility of malicious decisions is greatly reduced.

Besides, veCRV holders also enjoy the trading fee share, that is, the interests of the liquidity provider, trading party, token holder, and protocol are well coordinated. (What trading parties can take advantage of are the strong liquidity of the pool and the lower slippage.)

If it can go through the cold start period smoothly, Curve will run like an excellent positive wheel, and VE will play an indispensable role in it.

We can say that ve is a typical example that restricts human nature at the mechanism level and guides positive behavior.

1.2.2 Disadvantages

No model is perfect, and ve also has its points of criticism.

1) The rigid “lock-up time” is not friendly enough to investors:

The lock-up period is not only a sweet spot but also a point that discourages many investors.

Some people joked that 4 years in the crypto industry is equivalent to 100 years in other industries. Quite a few investors don’t want or are unable to lock their assets for such a long period of time. To develop further, it can focus on how to make it more attractive to attract a wider range of investors and offer flexible lockup periods.

2) Centralized governance:

Presently, Curve owns more than half of the governance rights (53.65%). The governance is quite centralized.

Note: Convex is a Curve-based liquidity staking and mining platform. CRV holders can stake CRV on Convex to obtain cvxCRV. The Convex platform will automatically lock the obtained CRV tokens on Curve, so as to get the veCRV tokens mastered by the protocol. It can be understood that cvxCRV tokens are veCRV that can be tokenized in circulation. That is, Convex places bets on Curve via certain rule designs and risk control, aiming to maximize rewards by taking advantage of the “scale advantage”.

https://dao.curve.fi/locks/1667977964

https://dune.com/queries/56575/112408

This aspect has something to do with the lockup time in 1). Non-big users may find it hard and don’t have enough motivation or ability to lock tokens for a long time. Besides, it is also related to Curve’s whitelist mechanism, which stipulates that smart contracts cannot participate in the DAO unless they are approved by voting (51% pass rate, 30% participation rate). It’s proposed to maintain the stability of the protocol. There are only 3 protocols that have obtained whitelist permission in history, namely, Yearn, Staking DAO, and Convex.

At present, voting rights are held in a few protocols, and, intuitively, they have no motivation to agree to allow more protocols to participate, which will lead to fierce competition. Curve war may be pushed to a new stage if new and innovative ways appear.

However, the whitelist is somewhat contrary to the fairness and openness that crypto has always advocated. If the early whitelist has played its role, it is debatable whether the permissionless ecology is more beneficial to Curve or even Convex in the long run. This Propsal (Remove Curve DAO Whitelist) is discussing whether the whitelist should be removed. It has been discussed widely among users. Users who insist on keeping the whitelist believe that the advantages of the whitelist still outweigh the disadvantages. If there are new protocols that are regarded as innovative enough, we should not worry about whitelists and block it because of voting.

Explain ve model and its innovation from its core mechanism

The Defi space is changing with each passing day, and constant iterations are made for the economic model. Curve is by no means the pinnacle innovation of ve mechanism. Different protocols have made different improvements based on Curve ve. Next, we will classify these innovations from different levels of the core mechanism, analyze the problems they attempt to solve, and see what we can weigh and optimize when designing ve model.

2.1 Liquidity balance

As mentioned above, the long lock-up period in Curve ve is an obstacle to many investors. The most intuitive way may be to shorten the maximum lock-up period. For example, Balancer sets its maximum lock-up period at 1 year. In subsequent models that are based on Curve’s own veCRV, there are also protocols that directly provide liquidity solutions. In addition, compared with Curve’s hard lockup, many relatively soft lockup methods emerge, trying to achieve a certain degree of balance between long-term lockup and liquidity.

2.1.1 ve token tokenization

Convex responded that ve’s low liquidity is caused by tokenization of ve tokens. Users deposit CRV on the Convex platform to obtain cvxCRV (only one-way conversion is supported in Convex). After staking cvxCRV, users will also get additional CRV (10% CRV mining revenue of Convex) and CVX rewards (Convex platform tokens) as well as trading fees in 3CRV sharing with veCRV holders.

Users who lock CVX (ie vlCVX holders, vlCVX=The Vote Locked CVX) can obtain governance rights of Convex and vote to determine the weight of each gauge.

As mentioned in 1.2.2 above, Convex holds more than half of the governance rights on Curve. Therefore, if they own the governance rights of Convex (vlCVX holders), they can indirectly control Curve in a relatively strong way.

Differently, the lock-up period on Convex is “only” 16 weeks + 7 days, which is much more flexible than Curve’s 4 years.

Furthermore, as the meaning of the word Convex indicates (the slope of the convex function is increasing), Convex is equivalently gathering the power of “retail investors” to obtain higher rewards. The increased annualized revenue adds to the attractiveness of Convex.

Both points also partly explain why Convex can quickly gain popularity after its launch (May 2021).

However, Convex, to some extent, will reduce the long-term consistency between veCRV holders and the protocol. This may be unavoidable after the liquidity is enhanced. Votium and other vlCVX vote-buying platforms have been derived from Convex, further increasing the governance level and complexity. (This article regards Convex as Curve’s functional L2, Votium as Convex L2, and makes a wonderful analysis of each Curve L2 protocol.)

On the other hand, attention should be paid to the decoupling risk between cvxCRV and CRV. The current ratio of cvxCRV/CRV is around 0.83, which has deviated from 1 for a long time. To this end, Convex is also launching a new solution, aiming to give more rewards to cvxCRV holders.

2.1.2 NFTization of ve tokens

In terms of lockup, the ownership of veCRV is non-transferable. In the VE(3,3)* model proposed by AC (the founder of Yearn), one of the important mechanism innovations is to make ve tokens transferable.

AC designed the locked ve tokens as NFTs, allowing the generation of multiple veNFTs when one account locks ve tokens multiple times. Each NFT can be accumulated to get the total amount of its account. After NFTization, ve tokens can not only be traded in the secondary market but also can further form a lending market in the future, thereby greatly improving liquidity. Since ve is linked to governance, liquidity providers and governance users (veNFT holders) are probably separated.

The question is: how should veNFT be priced? And if the locked ve tokens can be traded directly (even at a discounted price), where does the motivation for users to lock ve come from? Further improvements should be made at this point.

Note:

The VE (3, 3) model combines Curve’s ve model and OlympusDAO’s (3, 3) game model.

(3, 3) refers to the game results of investors under different behaviors. The simplest Olympus model includes 2 investors. They can choose to stake, bond and sell. As can be seen from the table below, when both investors choose to stake, the mutual benefit is the highest, reaching (3, 3), intending to encourage cooperation and staking.

https://olympusdao.medium.com/the-game-theory-of-olympus-e4c5f19a77df

2.1.3 Soft Lock*Exit Penalty

Compared with the mechanism of Curve which doesn’t allow users to exit midway after lockup, many protocols have extended the mechanism to allow for the midway exit but with an exit penalty. Penalties usually depend on the locked reward coefficient, which means that although users can quit halfway, they will also lose their locked rewards, increasing the opportunity cost of users quitting. Then, in some scenarios (such as short-term price fluctuations), users may choose to stick to the lockup.

Taking Platypus (a new type of stable currency AMM) as an example, staking PTP (Platypus’s native token) allows users to get vePTP and additional PTP rewards. A staked PTP can generate 0.014 vePTP/hour, and the upper limit of vePTP generation is 180 times that of depositing PTP (18 months).

The user can terminate the staking at any time, but with the cost that once the user releases any amount of staking, the vePTP accumulated over time will be cleared. This means giving the user the right to choose: whether to stick to the achievements accumulated in the past or to give up easily. It is conceivable that the superposition of users’ judgment and loss aversion will make their choices more subtle and diverse.

Compared to Platypus’s total zeroing rules, Yearn (a Defi revenue aggregator) is much gentler. In its upcoming ve model (based on YIP65 passed at the end of 2021), veYFI has a lock-up period ranging from one week to 4 years. Holders who choose to quit midway are subject to a variable penalty related to their remaining lockup time. If the remaining lock-up time is greater than 3 years, the penalty is 75% of the locked YFI, otherwise, it is calculated based on: the ratio = the remaining lock-up period / the longest lock-up period (4 years). This penalty fee will be rewarded to the remaining holders of veYFI.

GMX (DEX) is also a project that cancels rewards proportionally, and its staking rewards Multiplier Points will be burned proportionally.

In addition to penalty rewards, some protocols have chosen to extend the unstaking time. For example, Prism (derivatives protocol) stipulates a 21-day unstaking period.

Soft lock-up, to some extent, balances liquidity and long-term holding incentives. Redistribution of penalty fees to holders sticking to lockup also further incentivizes long-term holding behavior. However, the specific parameter settings (such as the longest lockup period, exit penalty, etc.) should be set according to the priority you want to achieve in the specific scenario.

2.2 Incentive Mechanism Setting

In this section, we will explain the incentive distribution mechanism in detail.

The sources of rewards for holding ve tokens mainly include:

1) additional token issuance rewards

2) an increase in the reward multiplier

3) transaction income share (such as handling fees)

4) proportional distribution of penalty fees and other benefits(if any)

2.2.1 Transaction sharing Incentives

First, let’s take a look at the sharing of transaction income, that is, whether ve holders enjoy the sharing and the specific sharing ratio. In theory, if users hold positive opinions about the long-term development of the protocol and the income of ve holders can be continuously bound to long-term income, it should increase users’ motivation in the long run. Taking Curve as an example, veCRV holders enjoy 50% of the transaction fee share of all pools, which are distributed in the form of 3CRV. This means that when the price of CRV drops, investors are prompted to buy CRV and buy a larger share of revenue at a cheaper price.

The ve(3, 3) model is further refined based on this. In other words, users only enjoy the handling fees generated by the pools they vote for. This new restriction prompts users to vote for the pool with the most liquidity and the most fees.

In terms of how to balance the income of veToken holders and liquidity providers, the specific sharing ratio depends on the consideration of the protocol designer. Curve adopts a 50% split method, while GMX protocol implements a sharing ratio of 3-7 between stakers and liquidity providers.

Furthermore, if tokens are used for transaction sharing, they will, to some extent, have stock properties. In U.S. law, the Howey Test is used to determine whether an asset is a stock. If it is classified as a stock, it will be brought under additional regulatory constraints. However, since it belongs to case law, the actual judgment should be made according to the specific case. Different countries also have different ways to judge.

2.2.2 Additional token issuance incentives

When issuing token rewards, Platypus (a new type of stablecoin AMM) has designed a three-pool model:

As for the proportions of additional issuance income, the three pools take up 20%, 30%, and 50% respectively (this weight can be adjusted later). AVAX-PTP (20%), pool 2, is the liquidity pool of PTP; the Base pool (30%) rewards liquidity providers, and the reward weight is proportional to its deposit ratio; the Boosting pool (50%) rewards ve holders and the reward weight depends on the deposit amount and the vePTP amount.

The current distribution ratio of the pool shows that the designer is tilting the rewards distribution towards vePTP holders (50% VS 30%). However, be that as it may, separating the Base and Boosting rewards makes sure, to a certain extent, Non-ve holders can get basic ARP.

What GMX adopts is that both Base and Boosting rewards come from the same pool. In this case, the income of non-ve holders may be greatly reduced, but to some extent, users are also encouraged to stake for the long term to obtain ve. It can be seen from this that designers can take whether to generate additional token issuance rewards for the same pool or not in advance into consideration to balance ve incentives.

2.2.3 Anti-inflation

Generally, protocol tokens are continuously issued. Therefore, locked tokens will also be subject to inflationary pressure if no action is taken. If the foreseeable inflationary pressure is too great and the tokens cannot be sold midway, it is obvious that users will be unwilling to lock. To solve this problem, ve(3, 3) proposed its optimization scheme, which is, specifically, as follows:

1) The weekly additional issuance will be dynamically adjusted based on the circulation

Assuming that the original weekly increase is 2M, if 0% of the current tokens are locked in ve, the weekly increase is still 2M, that is, 2M (1-0%); if 50% of the tokens are locked in ve, the weekly issuance amount is 1M, that is, 2M(1-50%); if 100% of the current tokens are locked, the weekly issuance amount will be 0, that is, 2M*(1-100%). In other words, the more tokens locked, the less the increase, thus reducing the impact of additional issuance on locked users by dynamically adjusting the increase.

2) ve lockers will obtain compensations proportionally

Assuming that the current total supply is 20M, the locked amount of ve is 10M, and the weekly increase of the week is 1M, which means that the supply increases by 5%, that is, 1M/20M. In order to ensure that ve holders are not diluted, their holdings will also increase by 5%, that is, 0.5M=10M*5%, in the (3,3) mode. The rest 0.5M of the weekly increase will be released as rewards.

If we analyze 1) and 2) comprehensively, we can see that the designer hopes to enhance its locking power by protecting ve users from being diluted.

Inflation is equivalent to taxation, which increases the user’s holding cost. Hence, if it involves additional token issuance, we also need to pay attention to the current circulation, maximum issuance and issuance speed when deciding to lock, while the designer of the protocol can refer to ve(3, 3), so as to make more detailed regulations on additional issuance and compensation.

However, on another level, it is also necessary to take the token incentive balance between the first entrants and the late entrants into consideration. If the rewards additionally issued are tilting greatly towards existing ve holders, it may be less attractive to potential new investors.

The new Defi Thena.fi has made further improvements on the basis of ve (3, 3), which limits the increase in holdings to 30%, and only provides partial dilution protection to ve holders, thus not bringing tokens under the over-centralized control of early ve holders. The main focus here is how token issuance can constantly fuel the long-term development of the protocol.

2.2.4 Whether it is sustainable at the Real Yield level

It is already a common practice for projects to promote cold start and holding through token rewards. However, after experiencing the radical market when defi protocol released a large number of tokens to attract liquidity in 2021, investors have also begun to calmly examine whether the APY is sustainable. The Real Yield is put forward, in which this pursuit of a stable income is reflected.

Real Yield: net revenue = protocol revenue - market value of additional tokens issued

This formula shows the intention that we should take the additional tokens issued to users as real costs, and deduct them from the protocol revenue to calculate the net revenue. It uses the net revenue to measure whether the protocol has the potential to achieve positive returns.

Defiman has done research on Real Yield. The results show that most of the leading protocols that implement token incentives are unable to achieve a positive real yield, not to mention unknown or early protocols that often adopt more radical token incentive measures. Its long-term sustainability is doubtful.

In the ve model, in order to give users enough long-term motivation with the limited protocol revenue rewards, the designers often choose to mainly focus on token rewards. Therefore, when incentives are set, it is worth thinking about how to reasonably set an incentive level that can not only give users enough incentives but also make the protocol sustainable. In other words, the transaction income is enough to cover this part of the reward expenditure, while the transaction itself can remain competitive (revenues such as handling fees are reasonable but not too high). But the balance between short-term and long-term is really difficult to achieve, and it also tests the designer’s pace.

2.3 Governance Mechanisms Tradeoffs

2.3.1 Governance rights for liquidity

One of the core purposes of the ve model is to achieve good and stable governance. And the key point is: who can hold the governance power and how can the governance weight of different people be distributed? Ideally, primary governance rights should be granted to investors who have the same interests as the protocol in the long run. Meanwhile, from the perspective of time dynamics, it is best that certain governance rights can flow among users.

If investors holding a large amount of veCRV control the protocol by controlling voting rights, and thus further block the entry of new investors (refer to Curve’s whitelist mechanism), it will be hard to change the entire protocol. After the governance rights are fully consolidated, there will be no so-called governance rights to a certain extent (small voters are unable to exert a substantial impact on the results).

Several points are involved here (without considering other specific restrictions and requirements):

1) No whitelist restrictions, permissionless interaction with the protocol (avoid monopoly, more open)

2) ve is designed to be non-transferable, avoiding direct vote-buying

3) The voting weight of ve decreases with time, and the governance rights can flow naturally

Taking vcDORA (a veToken used for funding community governance openly) as an example, the governance capability of a certain number of ve can be visually seen in the figure below (the curve in the figure below is an example, no need to focus on why it is presented in this shape). The area under the curve from the current time to the end time represents the sum of its governance capabilities. When the staking expires, the user’s governance rights become zero. However, if g(t) is a curve, it means that g(t) is good as long as it can accumulate points (that is, the purple area in the figure below can be calculated), and it is not necessarily a straight line of linear attenuation. This nonlinear governance power curve can also be a direction to be explored.

https://doraresear.ch/2022/09/16/vcdora/

2.3.2 Loyalty measurement

If we think that weight is used as a way to quantify the loyalty of ve users to the protocol (the weight determines the voting/reward distribution, etc.), there are currently two measurement methods: one is Curve, which depends on the remaining lock-up time (looking forward, veToken will decrease linearly during the remaining lock-up period); the other is to look at the user’s historical staking time (looking back, the longer the staking time, the more veToken accumulated, but there is an upper limit), for example, Platypus. To some extent, these two measures somewhat find a balance between new and experienced users.

If the second measurement of looking back dominates, it is difficult to achieve better advantages for investors who have staked for a longer period of time. However, whether users with a longer staking period are more likely to continue to insist on stable staking or not is a matter of opinion. On the other hand, the second measurement is often combined with a midway exit penalty mechanism (see 2.1.3). As a result, investors will suffer a certain opportunity cost, which may prompt them to continue to stake steadily.

The method of looking forward (mainly based on the remaining lockup time) led by Curve may bring about a more fluid weight distribution. Any new users who choose to lock for a long time will get enough ve, while old users will no longer own ve if their remaining lockup time ends and they no longer choose to keep locking. Therefore, if users want to keep getting a good reward multiplier, they need to update their lock-up time to keep their remaining lock-up time at a relatively long level. This also partly explains why Curve’s average lock-up period can reach 3.5 years.

2.3.3 Pros and cons of vote-buying

Vote-buying is a disputable topic. We’re more willing to believe that vote-buying is a neutral word, depending on the main demands of users in specific scenarios. After all, in real life, it is similar to another word “lobby”.

In some scenarios, users mainly focus on the rate of return, and the convenient vote-buying mechanism can improve the user experience. It is even possible to consider directly integrating the vote-buying reward mechanism into the protocol, thereby making ve’s income more attractive and increasing motivation to run for election. Small ve holders may find it difficult to commercialize their governance rights or need to pay high communication costs if there is no convenient way to participate in the vote-buying protocol.

Giant whales also face the risk of breach of contract if they conduct privately directly through OTC. The built-in bribery mechanism solves these problems related to trust. Even in Velodrome Finance (currently the DEX with the largest lockup volume on Optimism), bribery income contributes the main ve rewards (as shown in the figure below). Moreover, the built-in bribery mechanism will inevitably exert an impact on L2 derivative protocols such as Convex.

https://twitter.com/VelodromeFi/status/1616489024268402711

In some cases, if the focus is the authenticity of voting and it is expected that users can participate in voting as much as possible and reduce bribery as much as possible, ZK-based MACI voting (able to hide each people’s votes, but display the final voting results) can be used.

In this situation, firstly, the privacy of voters can be well protected, and secondly, the vote-buying party is unable to know the real voting situation, thus discouraging them from bribery.

It can be seen that these designs of the protocol are only tools which can be used in combination based on specific scenarios.

2.4 Attempts to separate/stratify rights and interests

Many protocols are also developing new token rights distribution methods on the basis of ve. They attempt to further split veToken to make different tokens owing different rights, so as to grant rights in a more refined way.

Essentially, the Convex mentioned above splits the income rights and governance rights of veCRV, and becomes more flexible: users who stake cvxCRV can get 3CRV rewards, and locking CVX to get vlCVX can affect the veCRV controlled by Convex, thus voting on governance decisions of the Curve pool.

2.4.1 Separating governance rights from reward increase rights

Ref Finance (the DEX of the Near ecology) proposed a new set of solutions in its 2.0 token design. Instead of obtaining a single veToken, liquidity providers can obtain veLPT and Love (Ref) (1:1) after locking their assets.

Among them, veLPT is non-transferable. Corresponding to users’ voting governance rights in ve tokens, users who hold veLPT can vote to determine the allocation of incentives. Love (Ref) corresponds to the liquidity (locked liquidity share) provided by users in ve tokens, and holders can enjoy the increased income that’s determined by veLPT voting. Love(Ref) is transferable, that is, if the user only wants to enjoy the right to vote, he can transfer the income increase right to others. However, when unlocking veLPT, users need to veLPT and Love(Ref) at a ratio of 1:1 in their account.

https://ref-finance.medium.com/ref-tokenomics-2-0-vetokenomics-on-testnet-c2b6ea0e4f96

We can see that different investors can choose the part they want most in the protocol according to their needs to focus on investing after the token rights are subdivided, achieving greater flexibility for all parties.

2.4.2 Right-interest separation

The Astroport (dex for Terra ecology) mechanism strikes a balance between tokens without lockup and governance rights and tokens with long-term lockup and high returns of Curve ve. It adopts a novel three-token mechanism with ASTRO, xASTRO (obtainable by staking ASTRO), and vxASTRO (obtainable by locking xASTRO, which decays linearly with time). xASTRO is transferable and allows users to enjoy certain governance rights and transaction shares (50%). In addition to the 50% transaction share, vxASTRO allows users to enjoy more governance rights and income increases (up to 2.5 times).

The merit of this model lies in that it accommodates the demands of both short-term holders and long-term holders. It makes the governance process more democratic and reduces the friction of participating in governance by also allowing users who hold xASTRO but don’t lock them to enjoy certain governance rights. At the same time, it gives vxASTRO holders excess rewards asymmetrically, which has fully encouraged users to hold for the long term. In this way, investors with small amounts but strong beliefs can amplify their influence via long-term holdings, which forms a subtle competition force with investors who have large amounts but hold for the short term.

By stratifying the income structure and implementing the specific distribution ratio, the protocol can effectively regulate the investor structure it wants to achieve and may attract a wider range of investors. If Curve implements 100% ve (only users who lock tokens can enjoy governance rights and income increase), we can set up an incomplete ve mechanism (offer non-lockers certain governance rights and income, such as 50% ve) by stratifying rights. However, the specific design depends on the actual considerations of the protocol designer.

Conclusion

The ve model coordinates the interests of all parties by using lock-in and revenue sharing. Therefore, participants are encouraged to contribute to the long-term healthy growth of the protocol. In terms of the economic model, ve is a great improvement compared to the previous one, having achieved great success.

Innovations made by different protocols based on ve emerge one after another, contributing iterative solutions from different dimensions. This process witnessed repeated adjustments made by each protocol at key levels of the mechanism according to their needs and priorities. The system designer can determine how much it can be adjusted and designed based on the specific situation. These tradeoffs display the subtleties. And the permissionless and composable features further encourage people to make constant innovations.

Furthermore, built-in mechanisms are the best way to interact with participants in protocol design.

We cannot believe that the assumed participants are all rational and can adhere to the (3, 3) principle, or are willing to agree with the interests of the protocol. In practice, we have to guide participants to make as reasonable a decision as possible through things such as locking up rewards, exit penalties, and opportunity costs. And by virtue of the rights stratification/separation, participants can select an investment plan that suits them best.

A well-positioned mechanism can guide the orderly development of the protocol, as well as provide long-term force and reasonable incentives, which is also the core of ve’s continuous iteration.

With veModel, we may also extend to better governance mechanisms. In the above table, we can get the final voting synergy result by assigning different weights to different factors. One of the yardsticks that judge a governance mechanism design is successful is that the mechanism can promote creating more value and bring in better benefit distribution.

However, when continuously refining the protocol, we must also restrain ourselves. AC once reflected on the design of Solidly in his article by mentioning that “Simplicity works decentralized, complexity doesn’t”. Yes, sometimes complexity doesn’t mean effectiveness. Also, the complexity should also go with the scene.

Is veModel necessary? Do ve tokens have to be tradable? Do non-ve holders also have to enjoy certain governance rights? There is no right answer to any of these questions.

Certainly, despite lots of innovations, it is ultimately necessary to test the effectiveness of ve model in the market. We can design the fineness of the veModel as needed, and constantly adjust the design according to the feedback. It’s expected to see more exciting veModel designs.