Sixteen years after Robin Hanson published a seminal paper on using prediction markets to inform governance decisions, his ideas have found a home in the Ethereum community.
The associate professor of economics at George Mason University and researcher at Oxford’s Future of Humanity Institute first coined the term “Futarchy” to describe a new form of governance that uses data from prediction markets to provide input.
By Michael del Castillo
Since the article’s original publication, Hanson has traveled the world teaching classes, pitching businesses and making appearances at conferences about Futarchy. But until now, no one has ever implemented the idea in the real world.
Recent moves in the Ethereum community have provided an avenue for these concepts to move closer to implementation. In interview, Hanson offered why he believes his simple idea finally found a fit in a distributed, autonomous world.
He told CoinDesk:
“The slogan is vote on values, bet on beliefs. What you need are discreet decisions and then you need an outcome that you care about.”
In the Futarchy model, market participants, or voters, are allowed to buy stock in an idea that may or may not take place over a certain period of time.
To follow Hanson’s frequently cited example, a public company could hold its chief executive accountable to achieving a particular stock price over a given period of time.
Those who believe in the CEO can invest in a “yes” token, thereby supporting the future success of the company and positioning themselves to get paid out if they are correct. Participants who don’t believe in this outcome can invest in a “no” token, and receive a reward if they are correct.
Effectively, authority holders set the terms by which their success or failure would be determined, and market speculators use their money to indicate the outcomes they believe to be the most likely – thereby setting an agenda of sorts within that governance structure.
When applied to a blockchain-based corporate structure, also known as a distributed autonomous organization (DAO), the prediction market concept could empower stakeholders to vote their outcome beliefs.
Some community members want to see this implemented in the near term, as the DAO concept begins to move past the drawing board.
Last week, Ethereum-based prediction market Gnosis published a private version of a proposal to augment The DAO – an organization aimed at funding Ethereum projects that has collected more than $150m worth of the cryptocurrency ether since its launch earlier this month – with a Futarchy-based prediction market used to vote on the pitches it receives.
Making contact
Hanson was first introduced to the Ethereum community back in 2014, when he says Ethereum creator Vitalik Buterin reached out to him with questions about Futarchy.
By the end of that year, Buterin had published a piece called “Introduction to Futarchy” that included a detailed description of how the concept functions, an account of Hanson’s CEO example, as well as five arguments against Futarchy.
Perhaps most notably, one argument included in the piece posited that wealthy entities could shape prediction outcomes by purchasing an outsized amount of “yes” tokens.
This initial contact with elements of the Ethereum community would go further, as Matt Liston, co-founder of Ethereum-based prediction market platform Augur, sought to bring in Hanson as a project advisor in late 2014. The arrangement was formalized in August 2015, after Liston’s departure from the company.
Improving DAO governance
A distributed organization built on a blockchain like Ethereum is governed primarily through the function of smart contracts. But in order for the smart contract to function, it first needs a form of input or data from the outside world – and that data is only as valuable as it is trustworthy.
Traditionally, information is provided to prediction markets like casinos in the form of a trusted third party. But in the decentralized world of DAOs that information could come from what is called an “oracle”, a source of information that relies on the “wisdom of the crowd” to formulate its answers.
“An oracle is an external actor which can provide information from the real world into the blockchain,” said Stefan George of Gnosis, whose company is building both a prediction market to help DAOs make decisions and and an oracle to help generate the data. “The blockchain itself doesn’t know anything.”
Users of Gnosis, for example, will be able to “sign” data which can then be used to help resolve events pertinent to other users’ desired decisions. For example, the data provided would give insight into whether a sports team actually won a game, of it a CEO actually delivered on the promises they made.
Finding the right product fit
At issues is the fact that The DAO, in particular, already has a method of governance.
Built from open-source code written by Ethereum-based startup Slock.it, The DAO has raised millions worth of of ETH based on a business model of allowing those who buy voters rights tokens to cast a vote on funding proposals they want to support.
But according to Martin Köppelmann, co-founder of Gnosis, the principles of Futarchy may offer a better route.
Köppelmann explained that the concept as its built into his company’s software provides a more accurate way for stakeholders to express confidence in outcomes, using a prediction market that rewards them proportionately.
Instead of casting a single vote for every DAO token they own, experts are empowered to vote relative to their confidence that they are correct. A virtual reality expert might have particularly strong feelings about a new immersive system build into Ethereum and with Futarchy would be able to express that confidence.
Köppelmann explained:
“In Futarchy, if someone has a very strong opinion and he would be very sure that he’s right then he could just use more money and therefore have a higher influence.”
While the system may seem easy to game by faking expertise, the long-term reputation of a voter can be damaged and lost resources would accumulate.
In theory, the risk of reputation loss amongst participants incentivizes honest behavior. But using rewards to dissuade corruption is a long way from actually preventing it.
The Ethereum Foundation Futarchy grant
On 25th April, Liston and Köppelmann won a $15k grant from the Ethereum Foundation to conduct three separate experiments on Futarchy using the Gnosis technology. The funds will be disbursed over a three-month period, with the project encompassing three experiments.
According to Liston and Köppelmann, the first two experiments are designed to test the impact of the Gnosis prediction market in a “scalar” setting, where market manipulators are incentivized to manipulate the value of a smart contract that is expected to pay out on a particular date.
The third experiment “gets a little closer to Futarchy” as Liston put it, testing what is known as the “should we hire this CEO” theory.
In this particular experiment, the vested interests of a CEO and his associates will be tested over time relative to the expected return. The hope, according to the two, is to figure out a way to prevent the system from being rigged.
Ultimately, Gnosis intends to seek to do business with distributed organizations, including The DAO, as it seeks to bring these governance concepts to fruition.
Asking the right questions
Hanson says that, in the 16 years since he first published his paper on Futarchy, the concept has seen mixed results on the implementation front.
“Often people get excited about the abstract but they don’t get the details right,” he told CoinDesk. “Getting the details right matters especially in these blockchain-based technologies where after it goes live you can’t just change them.”
With an eye to this degree of confusion, Hanson later published advice on how to navigate the trickier aspects of his concept. In conversation, Hanson suggested that special attention ought to be paid to precisely how predictive questions are phrased before they are put to the market for feedback.
Hanson:
“The key thing is you want to ask the question that you actually want the answer to, and you want to ask people who might actually be able to answer.”