Prediction markets are wonderful things. They apply the power of markets to elicit the best available entrepreneurial knowledge and insight, and they do so for things that ordinarily couldn’t be subject to market exchange.
A steel futures market, a market for the delivery of a definite amount of steel at a definite future date, is like any other commodity market. Entrepreneurs’ buying and selling decisions in this market reflect their best estimates of the future price of steel. What distinguishes a prediction market is that it allows the entrepreneurial process of valuation and exchange to be applied to any observable future event. Next year’s GDP cannot be exchanged as steel can be, but by creating a prediction market where people can bet on next year’s GDP, we create a market price reflective of the best available estimates of next year’s GDP.
I interviewed Zack Hess on my podcast, Economics Detective Radio, about his work to create decentralized, peer-to-peer prediction markets. Towards the end of the interview, Zack made the startling statement that he saw prediction markets as a step towards anarchy. The connection between prediction markets and anarchy is not obvious, so it is worth exploring just how the mere existence of prediction markets could undermine state power.
Suppose people like Zack are successful in creating decentralized prediction markets that cannot be shut down by governments. As was the case with the late intrade.com in the lead up to the 2012 US Presidential election, these markets’ prices will be a source of news stories and media commentary. Suppose that the public comes to trust these markets’ predictive accuracy.
This will change policymakers’ incentives. Today, in a world with few prediction markets, politicians can succeed in pushing harmful policies so long as the public can be convinced these policies are good. The social world is complex, and even people who dedicate their lives to studying the outcomes of public policies have great difficulties in assessing their outcomes. But in a world where prediction markets are both prevalent and trusted, a bad policy’s critics could easily demonstrate the policy’s bad effects. If policy A has bad effect X, its critics need only create two markets: First, a prediction market for X given that policy A is enacted. Second, a prediction market for X given that policy A is not enacted. The difference in prices in these two markets would lead to the common understanding that A increases the chance of X.
Supposing that X is sufficiently bad to make policy A unpalatable (as would certainly be the case if the whole point of policy A were to prevent X), the prediction market would destroy the credibility of any politician attempting to enact it. Thus, the political class would be constrained to the policies that could be justified by prediction markets. Even if the state’s authority were unchanged on paper, a political class so constrained would effectively go the way of the British monarchy: they would become symbolic figureheads, bereft of any real authority.
If you conceive of anarchy as simply the absence of rulers, then such a world would be anarchic. With the political class effectively dethroned, no small class of people could dominate the rest. However, if you conceive of anarchy as the absence of coercive institutions, the case that prediction markets will take us there is more difficult.
Ultimately, prediction markets can make the political system more adept at catering to majority preferences. The argument above assumed that the majority did not want X. However, if the majority favours robbing Peter to pay Paul, showing people that Peter is harmed will not undermine support for the policy. However, to the extent that coercive policies tend to harm both Peter and Paul, prediction markets will make this clear to everyone.
Ideological opponents can disagree on both means and ends. If prediction markets become widespread, and are commonly understood to be accurate, they will resolve much of the public debate about means. However, the debate about ends will rage on. If the public holds moral convictions that permit coercive intrusions into others’ lives, all prediction markets can do is expose those that fail to achieve their intended goals. A public that believes in the moral desirability of paternalism will still enact paternalist policies if prediction markets indicate that they will be successful. Prediction markets are no anarchist panacea. They cannot preclude the possibility of a majority using the state apparatus to coerce a minority.
However, if it turns out that the majority’s desired ends are best achieved by dismantling the state apparatus, prediction markets will certainly help bring this about.
 Note that this is not quite the same as Robin Hanson’s “futarchy” idea. Under futarchy, policy would automatically be determined by the betting markets on an aggregate measure of national welfare. This, on the other hand, is a kind of implicit futarchy acting through existing democratic institutions.
 This assumes, of course, that the harm can be measured and therefore bet on.