(This essay is a successor to my previous post Misconceptions and Value Capture About Governance Tokens , with some brief excerpts.)

As many DAO projects continue to mature, Governance Extractable Value (or Voter Extractable Value, hereafter referred to as GEV) becomes an interesting space to explore. You may have heard of bribes, which are a basic form of GEV. Contrary to the prevailing view that bribery is detrimental to a protocol, I hold the perspective that bribes, if leveraged prudently, can function as a cashflow that helps a protocol’s governance token capture more value. In this essay, I will delve into the rationale supporting this viewpoint, while also addressing exceptions such as the potential risks associated with takeover and parasitism.

Firstly, this essay will provide an overview of the significance and current state of the GEV landscape. Subsequently, It will then outline the goals of the GEV project, provide a comprehensive analysis of the design space within the GEV project, and derive from first principles the possible future development of this market. Through this exploration, I also aim to provide investors and project developers with insights into potential tradeoffs of design decisions.

Why GEV Matters

There are several reasons why GEV is important and worthy of investor attention. First, the GEV market is huge. If we consider DAO as the organizational structure of future organizations and on-chain voting as the primary form of DAO governance, then most protocols possess extractable value. In principle, GEV scales well with the number of DAOs and their governance activities.

Secondly, GEV cannot be eliminated. Governance tokens are intended to be a combination of two things — the governance rights of a protocol, and the economic benefits associated with that protocol. This coupling theoretically aligns economic gains with governance efforts and can incentivize participation. However, in practice, we must recognize that token holders have varying preferences for these two attributes, which is why there are bound to be some who will trade governance decisions for money. Furthermore, the composability makes the decoupling of these two attributes and transactions between them instant and effortless. For example, when holders of CRV (Curve’s governance token) stake CRV to Convex, Convex essentially decouples CRV into cvxCRV and vlCVX, the former representing economic interests in Curve and the latter representing governance rights in Curve. Economic interests are returned to CRV holders, while Convex accumulates governance power.

Thirdly, GEV brings cash flow that makes governance tokens valuable. Tushar of Multicoin Capital pointed out at its summit that the value of Layer 1 cryptoassets is the sum of three parts — commodity (providing utility), capital assets (producing income), and the nature of money. This concept actually applies to other cryptoassets, including governance tokens. In my earlier post, I argued that currently most governance tokens have little value, because 1) they have limited utility value (in fact, binding utility and governance right may create a misalignment between utility and token price), 2) most governance tokens don’t offer income to avoid regulatory scrutiny as securities, and 3) a generic buyback-and-burn does not equate to the distribution of income streams. Therefore, GEV in principle makes up the value of the governance token.

[Side note: I highly recommend reading my previous post, which argues that the value capture mechanism of many governance tokens is flawed and substantiates this hypothesis from first principles.]

Additionally, the timing is opportune for GEV capture. The operations of the initial batch of DAOs are approaching maturity (e.g., MakerDAO). The earliest GEV projects have emerged, sparking a “Curve War”. And new primitives in the GEV space are also being developed.

The Good, Bad, and Tricky of GEV

The concerns with bribery, as well as capturing GEV in general, are whether it is unethical and poses a risk to protocol security. To answer this question, first let’s consider a similar concept, Miner/Validator Extractable Value (referred to as MEV). Multicoin Capital has detailed, high-quality analysis of the situation and solutions here and here. In simple terms, MEVs can be divided into two categories — the good and the bad. Benign MEV is essentially transaction ordering within a block by builders and there is nothing unethical about it. However, malignant MEV is indeed unethical and potentially dangerous. It is often characterized by front-running/back-running or censorship.

Similarly, there are good and bad GEVs. Benign GEV occurs when the outcome is somewhat altered due to economic incentives offered to apathetic token holders yet no one is intentionally harmed. This alteration of voting results, by itself, does not pose a greater risk to the protocol’s security than the absence of incentives. A use case for benign GEV is to bootstrap new stablecoins on Curve. Malignant GEV, by contrast, often features “rug pull” — such as stealing from liquidity pools, collateral pools, or protocol treasuries — which is absolutely unethical and dangerous. A recent example of exploiting malignant GEV is here.

It is worth noting that the MEV-GEV analogy can be tricky — it has limitations in fully capturing the complexity of GEV, mainly in the following two aspects. First, actions of benign and malignant GEV may initially appear similar—more so than those of MEV. Secondly, governance decisions can shape a protocol (that is built on top of Layer-1 blockchains) over time, whereas benign transaction ordering is constrained by the L1 consensus mechanism. Thus, even a benign GEV can have uncertain long-term effects, which could be either positive or negative.

Objectives for GEV Projects

We’ve discussed above that GEV cannot be eliminated and can bring long-term effects, so it is imperative for GEV projects to have benign objectives:

1. Form a good dynamic with the protocol/DAO they are based on (hereafter referred to as “base protocol”)
2. Aim for long-term, sustainable value capture rather than short-term exploitation
3. Promote participation and facilitate coalition in decentralized governance
4. Ensure transparency at every stage of GEV extraction

I want to emphasize that objectives #1 and #2 require a win-win mindset from project founders and involve a trade-off between maximizing benefits and making sound decisions. It is important to note that even an initial win-win scenario may transition into parasitism, which I will explore in detail later in this post.

In GEV projects, coalition-building (#3) is more important than in the native governance of the base protocol because concentrated votes contribute to the efficiency necessary for effective GEV capturing. Additionally, transparency (#4) acts as a safeguard against harmful GEV practices.

Basic Models of GEV Project

Let’s examine some existing projects before exploring the design spaces in the next section. I will illustrate using the Curve ecosystem where the earliest GEV projects emerged. Curve allows CRV holders (usually LPs) to vote on which pool rewards should be distributed to. In order to secure liquidity for respective pools on Curve, projects began employing strategies to incentivize CRV holders. Two prevalent models of bribery emerged:

• bribe.crv.finance’s model, which I also call job board model. Projects looking to increase liquidity through Curve can post on bride.crv.finance indicating incentives to let veCRV holders know.
• Votium’s model, which I also call vote allocator model. vlCVX token holders can simply delegate their voting power to Votium, and Votium will allocate these tokens to vote for pools that offer the highest incentives.

Comparing these two models, I would conclude that the vote allocator model will gain popularity and create a flywheel effect, while the job board model will gradually be obsolete. Although many bribe givers and takers continue using bribe.crv.finance because it was the first to emerge, it is nothing more than a “job board”, which does not manage any risk nor add much value to the ecosystem. According to Multicoin Capital’s post Protocols Don’t Capture Value, DAOs Manage Risk, such a “job board” is very likely to fork. In contrast, the vote allocator model has several important actors, each tightly integrated into the ecosystem and managing some risks, as illustrated and explained below:

• Governance token stake pool (e.g., Convex): Stake pools serve to amass governance tokens to condense governance rights and create conditions favorable for extracting GEV. In general, staking pools assume risks to provide supplementary yield, create value by coordinating proposals that generate GEV, and extract some fees.
• Vote allocator (e.g., Votium): The vote allocator automates voting and aims to receive the highest incentives per token. Vote allocators undertake risks to vote for high incentives in a dynamic setting, create value by saving holders’ time and effort, and can levy fees.
• Reward manager (e.g., the Union by Llama Airforce): The reward manager collects all rewards from all users, and swaps them for the most common tokens, which are further put into an autocompounder until claimed by original governance token holders. In contrast to MEV, which is typically delivered as the L1 native cryptocurrency or close variants like ETH and stETH, GEV in DeFi can consist of multiple tokens from various projects. Claiming these rewards is not only sensitive to rates but also gas intensive. As a new actor emerging in the GEV ecosystem to solve this problem, the reward manager takes on the risk of volatility of swapping and gas, creates value by saving holders’ costs, and thus is able to charge for services.

The GEV-related interactions between these actors (① — ⑦ in the chart above) create a closed loop that tightly integrates various actors and achieves a flywheel effect.

Exploring the Design Space for Capturing GEV

The design space for capturing GEV is vast. In this section, I will delve further into the design decisions pertaining to a more generalized GEV ecosystem, which might evolve from the specifics of Curve discussed above. Taking a top-down approach, I will only focus on the key design choices that hold the most significance.

Generalist vs. Specialist

The first high-level decision to make is whether a GEV project will be a generalist or a specialist. A specialist project serves only one or a few base protocols, such as Convex for Curve or Aura for Balancer. A generalist project serves multiple base protocols, such as Hidden Hand, supporting 9 base protocols as of the time of writing. This decision holds implications for investors as well — While the GEV market is undoubtedly substantial, if the majority of GEV projects are specialists, each project can only capture a small portion of the overall market, limiting their upside potential.

My view of this design decision is that GEV space should have both specialists and generalists, with top base protocols dominated by specialists; but I expect that more projects will transition from specialists to generalists as the space evolves. For example, Pirex began as a GEV project for Convex but later added support for other protocols such as GMX (a decentralized perpetual exchange). This trend is favorable for investors and GEV takers.

There are several reasons to support my perspective. From the standpoint of GEV project founders, they are motivated to capture a larger portion of the GEV market, so they’re likely to expand to other base protocols and unlikely to reinvent a wheel for every base protocol. On the user side, some aggressive bribe givers and takers look for the highest bribe efficiency and yield respectively, and will prefer generalists to reduce migrations. This has been observed in cases where users migrate from Convex/Curve to Aura/Balancer, as there are currently more GEVs to extract in Balancer than in Curve.

Ecosystem Structure and Actors

Another crucial aspect of the high-level design decisions involves the construction and components of the GEV ecosystem. Specifically, how actors — stake pools, vote allocators, and reward managers in Curve’s case discussed above — are generalized around other base protocols and how they will evolve over time.

Let’s begin by considering governance token stake pools (alt-Convex). Some doubt whether they are needed for a generic base protocol, stating that Convex emerged before the prevalence of bribery and primarily aimed to circumvent Curve’s locking restrictions; if a base protocol does not require locking or staking or has minimal restrictions, the need for a stake pool might not be supported. However, I believe that stake pools are indeed necessary because they have a critical function — serve as an entity to amass governance power, facilitate coalitions, and provide additional yield. To draw an analogy, stake pools in L1 blockchains initially aimed to lower the threshold and resolve technical difficulties (e.g., in early Ethereum), but now they are needed to gather more staked tokens and generate higher yield through coordinating MEV. The same concept applies to governance token stake pools.

Next, let’s examine the role of the reward manager (alt-Union), which manages proceeds for GEV takers. I anticipate that reward managers are also necessary for a generic GEV ecosystem and will become increasingly important as the crypto space evolves. Imagine a future where you receive various cryptoassets as GEV rewards from DeFi protocol governance or even hundreds of NFTs as GEV rewards from Web3/creator economy project governance. It wouldn’t be surprising if GEV reward managers eventually integrate with crypto asset management projects (e.g., dHEDGE) or even crypto wallets.

Finally, we have the vote allocator (alt-Votium). While I acknowledge its crucial function — automating voting to receive the highest incentives — it finds itself “stuck in the middle,” facing integration pressures from stake pools and reward managers. For instance, Bribe Protocol, in its pilot with Aave, integrates the selection of the highest bid into its pool services. Additionally, in the long run, to ensure sustainable GEV extraction, I expect the allocation algorithm to evolve and consider factors beyond pure economic incentives.

Investors should also be concerned about which actor dominates a GEV ecosystem, as the dominant actor holds the potential to capture a larger portion of the GEV in the value chain. Vote allocators can be ruled out for the reasons mentioned above. Governance token stake pools are highly likely to dominate the ecosystem around a base protocol since the largest pools inherently accumulate governance power, which directly affects GEV accrual. Unless reward managers can establish themselves as sticky and defensible, they are not as advantageous as the largest stake pools in the race for dominance.

It is also important to note that different actors in the ecosystem have varying positions along the generalist-specialist spectrum mentioned earlier. Stake pools tend to lean towards being specialists, with their defensibility stemming from their shares of governance power around base protocols. On the other hand, reward managers are more defensible when they are more versatile and composable, tapping into multiple pools and managing a variety of crypto assets.

Own Tokens and Parasitism of GEV Projects

The final major decision to consider is whether GEV projects should issue their own tokens. This decision holds great significance for investors as it directly impacts how they monetize GEV. Normally, investors only need to hold the native tokens of the base protocol, which GEV is directly accrued to. However, if tokens from the GEV superstructure are parasitic on native tokens of the base protocol, then holding tokens from GEV projects becomes more attractive for investors.

Reward managers don’t have to issue their own tokens, as their usual fee structure involves taking a percentage from the cryptoassets they manage. Regardless of whether they choose to issue tokens or not, they do not have a parasitic relationship with the base protocol, because they solely deal with GEVs that are already captured and do not affect how or to whom GEVs are accrued.

In contrast, stake pools will almost certainly issue their own tokens. While the initial intention behind these tokens is to represent a deposit (similar to staking ETH in Lido and receiving stETH) or to offer incentivizes (similar to providing liquidity to Uniswap and receiving UNI), if these new tokens confer governance power over the base protocol, they can be considered parasitic to the native token of the base protocol. Parasitism, in this context, refers to the capturing of GEVs that would have otherwise accrued to the native tokens. For instance, currently, 1 CVX (Convex’s token) is equivalent to 5.16 CRV (Curve’s native token) as measured by voting power (source), and CVX/CRV price ratio is even higher at 5.66 (source), indicating a parasitic relationship.

It’s important to note that the boundary between symbiosis and parasitism is not absolute and involves tradeoffs. A popular GEV superstructure can generate traction for the ecosystem of the base protocol. I find a combination of parasitism and symbiosis acceptable, as long as GEV stake pools enable a transition for investors from holding native tokens to obtaining GEV pool tokens by allowing a native conversion between these two. I’m also looking forward to how things will develop if say, one day, a GEV stake pool gains control of over 50% of the voting power and decides to stop recognizing the base protocol’s tokens. This scenario might prompt future base protocols to adopt a crypto version of poison-pill defenses, such as issuing additional free tokens to current holders to dilute the pool’s share as it approaches 50%. Nevertheless, I am optimistic that through a few such conflicts, effective mechanisms will emerge.

Conclusion

The market of GEV is large and scalable as DAOs continue to develop. GEV plays a crucial role in boosting the value of governance tokens, particularly since many of these tokens currently lack revenue streams and utilities.

The goals for GEV projects revolve around ensuring transparency, facilitating coalition-building, and integrating effectively with the ecosystem surrounding their respective base protocols. These involve tradeoffs between extreme exploitation and sustainable practices of value capturing.

Through an examination of two existing models of GEV projects, the vote allocator model emerges as the prevailing option due to its tight integration with the ecosystem and its ability to manage certain risks. As the GEV space evolves and expands to include other protocols and DAOs, stake pools are likely to dominate the ecosystem and issue their own tokens. This dynamic can lead to either parasitism or symbiosis, or a little of both. In contrast, reward managers are more likely to offer generalized solutions that effectively serve multiple protocols.

GEV presents an exciting opportunity for investors, particularly in two prominent verticals: dominating stake pools for governance tokens and versatile reward managers for GEV proceeds. Moreover, several interrelated verticals can be considered, including generic cryptoasset management, algorithms and tools for vote allocation, and defenses against DAO takeovers. These aspects contribute to the broader landscape of GEV and offer diverse and healthy prospects for investment and engagement in the emerging decentralized governance ecosystem.

Thanks to Kyle Samani and Vishal Kankani for providing feedback on this essay.

About the Author

With recent experience in both investing (as an angel investor) and operating (as a product manager for a startup), Haydon Luo [website, LinkedIn, Twitter] has developed well-rounded skills for both sides of the same coin. Previously, Haydon began his career at AECOM, a multinational infrastructure consulting firm.

As a dedicated fitness enthusiast, Haydon is also committed to the blockchain space through volatile cycles for many years.

Haydon holds a BEng from the Hong Kong Polytechnic University and is on the verge of completing an MBA from the Carlson School of Management at the University of Minnesota. He speaks English, Cantonese, and Mandarin Chinese.

DISCLAIMER

The analysis in this document is for general information purposes only. While I endeavor to keep the information up to date and correct, I make no representations or warranties of any kind, express or implied, about the completeness, accuracy, reliability, suitability, or availability with respect to the document or the information, products, companies, or related graphics contained in this document for any purpose. Any reliance you place on such information is therefore strictly at your own risk.