Apply Here

Background

Decentralization is a key defining factor to the creation and maintenance of systems adjacent to cryptocurrencies, including but not limited to the entire web3 and DeFi ecosystem. Such systems often need structures that support their (decentralized) operation, for example by deciding in a permissionless way to upgrade some part of mutable code or parameters that the code uses to execute. It is then desired, if one were to support the idea of decentralization, that such structures also be designed and operate in a decentralized way. These decentralized governance structures are alternatively known as DAOs (Decentralized Autonomous Organizations). DAOs have control of treasury funds and are responsible for protocol upgrades and any other decisions related to protocol-specific parameter setting, so it is desirable for them to maintain decentralization while being resilient to, and able to recover from, attacks, as well as incentivize prudent long-term sustainability.

In many cases, it is not immediately clear how to judge whether a network or protocol has achieved this decentralization promise, or even in what way this is measured. It is important to study the regime of adequate decentralization in decision-making because the impact on the corresponding network or protocol effectuated by the votes of DAOs have critical impact on the health and behavior of that ecosystem.

DAOs are already operating at scale with large financial impact, but there has not been adequate research and formalization to measure and ensure their proper functioning.

Problem statement

Three issues are commonly identified with the operation of DAOs. First, most of the time the voting token distribution is not decentralized “enough”. Second, even when it purportedly is, there is a lack of active participation of a sufficiently decentralized set of stakeholders in votes, leading to voting outcomes that may not be representative of the (weighted) majority of the population. Third, incentive alignment between governance participants and the rest of the stakeholders or protocol/system participants is largely uncharted territory, as it’s not always obvious that self-interested voters will act to the “long-term benefit” of the system or protocol.

Multiple solutions have been devised to the problem of ensuring decentralized power distribution, one of which is to attempt to airdrop governance tokens to users that might potentially be stakeholders in the ecosystem [ex. Arbitrum]. In this case, sometimes protocols might try to disseminate such tokens by allocating a portion of them as well as call options on them to sophisticated financial marketplace actors [] that perform market making on them; the market makers then sell these into the open market without risk, and anyone interested in the protocol might be incentivized to partake into buying some of them, proportionally to their interest. This is, of course, the intended functionality, which may, due to incentive misalignment between the various participants in these markets, not be realized at all.

Another proposed solution to this first problem above, which seems to be more pre-eminent, is the “delegation” of votes: during a vote, some subset of governance tokens are (voluntarily or involuntarily) allocated to (presumably different) voting parties (commonly, addresses) who are free to direct the vote as they wish but do not otherwise obtain the token’s control. However, some very simple delegation mechanisms have been found in theoretical work [Gersbach et al., 2021] to be unfavorable to discovering equitable (democratic) outcomes. Such work is closely related to what is traditionally known as “liquid democracy” in the game theory literature.

The need for mechanisms to achieve decentralization in voting, showing the separation of token holders and actual voters, is clear in the case that the token distribution might be skewed: the tokens might represent some kind of valuation in the system (protocol or network), and thus, might be connected to financially-supporting actors which already have acquired them through some means. Thus, even if the governance tokens are not distributed uniformly at random, or equitably in a population, study of mechanisms that attempt to enforce decentralization constraints in governance decisions is still noteworthy —albeit to a limited extent, because of pseudonymity, or with appropriate assumptions of independence between (subsets of) voting addresses.

Proposal objectives

In general, of interest in this proposal is theoretical and/or empirical work related to the above raised issues and mechanisms that can shape how governance systems operate; for instance, an interesting problem would be to examine a variety of voting rules, and token distributions to achieve “desirable” properties. The overarching outcome of such research should be to provide actionable insight into how the various game-theoretic mechanisms at hand play out, as well as articulate clear guidance for the future operation of DAOs.

A successful project will result in a technical report or research paper achieving the above outcome.

Theoretical research will almost certainly need to involve making assumptions to whatever extent might be essential, and will be closely related to the field of mechanism design, as well as broader game-theoretic analysis of existing or proposed mechanisms.

In Places to Start, we point to a number of places where related questions have been explored.

Places to Start

One approach is to consider the various voting rules that have been implemented or proposed in the literature, or token distributions as they have been performed in the past or in the way that they are planned to be executed, and extract game theoretic (i.e., theoretical) properties of those. This could also include some portion of empirical research, quantifying metrics under which various systems or votes fared or did not fare well, but should go beyond just observational insights, and into actionable insights and robust future guidance. This is why we are suggesting that formal theoretical studies should be the primary scope/objective of this request for proposals.

Resources

• Papers
  • Token-Weighted Crowdsourcing [Tsoukalas and Falk, 2020]
  • Vote Delegation with Unknown Preferences [Gersbach et al., 2021]
  • Governance Extractable Value [blog post]
  • Vote Delegation and Misbehavior [Gersbach et al., 2021]
  • Voter coalitions and democracy in Decentralized Finance: Evidence from MakerDAO [Sun et al., 2022]
  • “Liquid democracy” game theory and mechanism design papers:
    • Flexible Representative Democracy: An Introduction with Binary Issues [Abramowitz and Mattei, 2019]
    • A Contribution to the Critique of Liquid Democracy [Caragiannis and Micha, 2019]
    • Weighted Representative Democracy [Pivato and Soh, 2020]
    • Liquid democracy: an algorithmic perspective [Kahng et al., 2021]
  • Governance models in stablecoins:
    • Stablecoins 2.0 [Klages-Mundt et al., 2020] (capital structure models of governance economic security in Sections 4 and 5)
    • Decentralized Governance of Stablecoins with Closed Form Valuation [Huo et al., 2022]