Blockchain Governance must move beyond voting
There is conversation that new voting mechanisms will revolutionize blockchain governance.
Given the struggles that many projects are having designing effective governance systems, the idea of implementing an innovative voting mechanism can seem like a panacea for the challenges of decentralized decision-making.
An example of a voting mechanism that is getting a lot of attention is quadratic voting (QV). First introduced by Eric Posner and Glen Weyl in 2013, QV allows voters to cast multiple votes in support of their preferred option in an election, and thus express their intensity of preferences over the options presented.*
In QV, a passionate minority that strongly prefers one option can prevail over a lukewarm majority by casting multiple votes for their preferred option. This prevents the so-called “tyranny of the majority,” in which the majority wins by number of bodies alone. The Ethereum community has embraced QV, and there are serious discussions of how it can be introduced widely on that protocol.
Quadratic voting adds to the decades-long tradition, dating back at least to Arrow (1951), of economists studying social choice theory, or systems by which groups can engage in collective decision-making.
While QV is a great example of innovation in social choice, neither it nor any other single voting mechanism is a panacea for blockchain governance.
The blockchain community must expand its thinking on governance beyond searching for the “perfect” voting mechanism.
A quality governance process needs to be custom-designed to best fit both the environment in which it is used and the decision it will be used to make.
- Who should participate in each decision?
- Is expertise required?
- How many votes should each person get, and why?
Building the governance system to fit a specific application is incredibly important if the results of the system are to be useful to the community.
In some cases, this may mean that a single platform will have different governance systems for different decisions that the community needs to make.
The challenges that Ethereum is facing implementing QV are a great illustration of why “peanut-buttering” a voting mechanism across environments and decisions can be a bad idea.
First, many commentators have already noted that QV is difficult to implement in an anonymous environment, such as a permissionless blockchain. QV allows each voter to cast multiple votes, but the value of each incremental vote decreases. The first vote that one casts is worth more than the second, which is worth more than the third, and so on. In an environment where it is very costly to have multiple identities (such as federal elections), one has no choice but to cast all their votes under the same identity. But if creating new, anonymous identities is almost costless, rather than have my votes decrease in value, one would instead create a new identity for each vote they want to cast.
Absent a credible method to uniquely identify each user and prevent multiple identities, QV reduces to voting by stake or another standard system.
Second, blockchain environments have different possible outcomes than other settings where governance systems are used. For one, a blockchain environment is especially unique due to the possibility of hard forks.
In the presence of nearly costless hard forks, the tyranny of the majority, which quadratic voting exists to solve, is less of a concern.
This means that if a minority of users don’t like what the majority chooses to implement, they can almost costlessly leave and start their own chain.
The question then becomes how do we design governance to ensure that only socially beneficial hard forks occur?
In our recent paper, Blockchain Upgrade as a Coordination Game, we show that quadratic voting can perform better than majority rule in some situations and worse in others.
Overall, we show that quadratic voting does no better than majority rule at coordinating the community on socially-beneficial hard forks.
As we discuss in our paper, an effective governance process to manage hard forks will need to also include coordination mechanisms, a defined policy proposal process (to determine what decisions get brought to a vote), well-defined information systems, and other elements.
While new voting systems such as quadratic voting have many applications both within and outside blockchain, their uniform application will not solve the governance challenges faced by blockchain today.
Moving forward, blockchain founders will need to think comprehensively about what constitutes governance and how these tools can be shaped to support their specific platform.
Many thanks to Bhaskar Krishnamachari for posing the questions that led to this post.
* We have simplified QV for the sake of exposition. For a detailed explanation of the mechanism and the math behind it, see Posner and Weyl (2015) and subsequent writings by the authors.