The Road to Futarchy
Futarchy has the potential to be an effective means of making objective governance decisions. It leverages markets to predict how decisions will impact metrics deemed valuable by a community. In the words of the concept’s progenitor, “Democracy would continue to say what we want, but betting markets would now say how to get it.” To better understand futarchy and to fully appreciate this article, see the full original post. DAOs such as dxDAO, MakerDAO, Aragon and others are coming online and making significant decisions through on-chain voting. Could futarchy improve the decision making abilities of these DAOs? Before a live DAO with millions or even hundreds of millions of dollars at stake decides to try it out, their participants will want to understand futarchy better. We must first demonstrate the viability of futarchy. Level K plans to launch futarchy signaling markets as a way to trial the concept. The signaling markets will run alongside onchain votes and show how the market views the impact of the vote outcome on a given metric. Examples of this could be predicting the price of ANT token if a governance proposal is accepted, predicting the future DAI/USD rate if the Maker stability fee is modified, and predicting whether the market capitalization of magnolia will go up or down if the DutchX liquidity fee is increased. We will be able to compare the results of the signaling markets to the results of onchain voting and also see how futarchy predictions pan out over time.
In this post we define Level K’s roadmap for bringing futarchy online as a viable governance tool for DAO’s. The roadmap is broken into efforts, prioritized by the order we intend to work on them, though some or all efforts may be carried out concurrently.
Level K Background
Level K, founded in 2017, helps teams build decentralized applications. In mid-2018, Level K worked with Gnosis to implement a futarchy curated registry on Ethereum using Gnosis prediction market and LMSR automated market maker smart contracts. Later in 2018, Level K received a grant from Aragon’s Nest program to implement a futarchy app for the Aragon platform, using the futarchy smart contracts developed with Gnosis. Level K is currently working on an oracle manager app for the Aragon platform. The oracle manager and futarchy apps will be used to launch futarchy signaling markets for Aragon Network votes.
Effort 1 — Oracles
The first item on our roadmap is an oracle framework for our futarchy markets. Since futarchy depends on prediction markets, we’ll need dependable oracles to determine the real world outcome the prediction markets are trying to predict and to thus set the payouts from the prediction markets. For example, if a prediction market is for the price of X token on date Y, we need an oracle to tell us the price of X token on date Y so that participants still in the market on date Y can redeem their rewards based on the difference between their prediction and the outcome. It is important that these oracles cannot be easily manipulated, lest they become a way to attack the prediction markets and manipulate futarchy decisions; they should be robust and decentralized.
Level K intends to use token price as the success metric for the first signaling markets we launch as token price is widely applicable to crypto projects. To support this, we are actively developing price feed oracles.
A robust, decentralized price oracle has the potential to be a public good for the Ethereum ecosystem and blockchain ecosystem as a whole. In evaluating sources of onchain data for our price feed oracles, Uniswap and DutchX stood out.
Uniswap can currently serve as an on-chain price oracle for many ERC20 tokens. It provides accurate market prices even when liquidity is low. However, it has a weakness as an oracle because it is possible for miners on Ethereum to sandwich a price query between a large buy and large sell inside one block. There is a proposed issue on the Uniswap repository to address this.
The DutchX can also serve as an on-chain price oracle. It is more resilient to price manipulation by miners. However, it requires more liquidity to operate. The DutchX can be a useful oracle for token projects with sufficiently high market cap.
Level K is currently building Uniswap adapters for our price feed oracle which aggregates and calculates the median of price data. These smart contracts will be used to set the outcome of our first signaling markets and will provide some mitigation to miner attacks related to short term price manipulation. Future integration with DutchX will also help mitigate price manipulation.
Effort 2 — Signaling Market Launches
The first signaling markets we intend to launch will be alongside Aragon AGP votes. The interface for the signaling markets will be an Aragon DAO and we will use the Aragon CLI along with an Aragon DAO kit to launch them.
The DAO kit will contain 4 Aragon apps, the oracle-manager app and futarchy app that we have developed, along with the voting and token-manager apps.
Further down the line we may develop a CLI tool or even a user interface to facilitate signaling market launches.
Effort 3 — User Interface
The mechanics of futarchy are complex and can be difficult to understand. Since we want futarchy to be a tool for DAOs, making futarchy accessible and easy to use has been a principal goal of our design efforts.
If you would like to be involved in user testing, you can check out the Aragon futarchy app on Rinkeby or reach out to us at email@example.com
Effort 4 — Permissionless Market Funding
For a market to exist among rational actors, there must be some intrinsic demand for the market. In futarchy markets, this demand comes from the desire of a group to make a decision. The group that stands to benefit from making good decisions, as defined by the impact on their success metric, stands as the natural party to provide initial liquidity for the futarchy markets. Since it is possible to profit from trading fees, third parties with adequate expertise may also choose to fund initial liquidity.
Initial liquidity is needed to seed the market. The more funds used to initialize a market, the deeper the markets will be and the less slippage will occur with trading. This article has a good explanation of the mechanics of the LMSR.
The starting point of the price is the midpoint between the upper and lower bounds of the scalar market, and currently, the creator of the market needs to set the initial price point via lower and upper bound parameters. We are looking to make this more flexible and also at using previous resolution prices to inform initial price setting.
In the current implementation of futarchy, the initiator of the futarchy markets provides the initial liquidity. Our goal for future versions would be to make the ability to provide the initial liquidity permission-less, allowing anyone to participate. One idea for this is to have a liquidity gathering period where participants can deposit funds and get token in return representing their portion of the liquidity.
Effort 5 — Vote Delegation
Signaling markets are a first step towards demonstrating the power of futarchy. They can teach us about futarchy without having a direct impact on voting. If they prove promising, futarchy markets could eventually be integrated into DAO governance. A first step in that direction could be to allow vote delegation to futarchy markets. Futarchy might eventually be used to directly make a DAO’s decisions, without token voting being involved.
For DAO’s currently using voting schemes to make decisions, integrations could be built to allow voting power to be allocated to futarchy markets. For example, imagine a network with 1000 tokens that uses token weighted voting to make decisions adding the ability to delegate votes to futarchy. If one party controls 40 tokens and delegates their votes to the futarchy markets, and the futarchy markets are favoring option A over option B, then those 40 tokens have effectively voted for option A.
Effort 6 — Research on Automated Markets
The scalar prediction markets used by our implementation of futarchy leverage Logarithmic Market Scoring Rule automated markets to facilitate trading. Automated markets guarantee liquidity to users, even in thin markets. The LMSR is an attractive scoring rule to use with prediction markets, providing a bounded loss for liquidity providers and allowing markets of varying depth to be initialized.
In the future we may explore different market makers that offer different depth and loss characteristics, or the possibility of setting up dynamic scoring rules that respond to the trading activity of the markets, increasing the depth as the demand for trading increases.
Level K is excited about decentralized governance and focused on making futarchy a reality. Follow us on twitter @levelk_io and if you would like to help us test futarchy, reach out at firstname.lastname@example.org
Thanks to Luke Duncan and Louis Giraux for their review.