Prediction Markets, Betting on the Death of Loved Ones

I want to argue that price is a great way to aggregate information. This can be shown in classical stock and commodities markets. Prices can also incentivize people to show their preferences in an election, or predict whether academic results are replicable. These prices are revealed in a prediction market, where futures contracts can be bought and sold on any future event. It’s like a bet on an event of your choice with odds that you post, reflecting your predicted probability of that event happening. I will then expand that notion explaining why this prediction market system would be great for social markets: investing into marriage or hedging against the death of loved ones.

Prices convey information. They convey information on demand and supply. More specifically on the elasticity of demand and the factors of production of supply. But they also bear information on what the future might hold. The price of a stock is determined by the market’s equilibrium prediction of how strong future discounted profits will be. Prices are a great mechanism to incentive knowledge aggregation. If you have reason to believe that something is underpriced you will buy it, thus increasing demand, thus increasing price. If you think something is overvalued you might sell it (or short sell it: borrow it, sell it, and buy it back at a later stage). This selling will result in a marginally decreased price in the market. Prices are a great mechanism to incentivize people to share knowledge. If you think you know better than the current equilibrium you can make money and by doing so add your information to the equilibrium. This leads to an efficiency in which prices convey all information in the market, the so-called efficient market hypothesis (EMH by Farma & French).

This system works well in the realm of stocks, commodities, bonds, and others, in which a liquid and accessible market is available. But what about other domains, where aggregate knowledge is vital? Two main examples are polls or national elections and the replicability of academic studies. The gold standard for both crucial predictions was asking people. In the example of polls, a national representative sample is asked who they would vote for on election day. Adjusting for a few sampling biases and inaccuracies, this would result in the polls we love and need, to plan our future with future governments. In the example of academic papers, experts were asked how likely it is that certain results can be replicated. This will indicate whether vital medical research will thrive or not. In both cases, asking people is a bad idea.

As I have pointed out in Lying to Ourselves is the Optimal Strategy and Why I don’t see a Therapist that we cannot trust what people say. Instead, we have to give people a real and salient incentive to answer questions truthfully. The best way to do that is by creating a market. As mentioned above this market can aggregate knowledge by incentivizing people to share their information truthfully. It can be shown that creating a liquid betting market on who will win a presidential election (mostly) better predicts outcomes than national representative polls. This makes total sense to me. I don’t want to tell other people who I will vote for. And if I had a cute hipster pollster ask I’d probably say “the greens” rather than “the neoliberals”. In 2016 people were reluctant to tell people they would vote for the Donald, resulting in every poll indicating that Hilary will win. This is a clear case of people lying to pollsters. But would they still lie if money was to be made? People were well aware of lying and knew that others might be too. If you could now earn money with this knowledge, you would reveal your preference. If you see odds reflecting a 90% chance of Hilary to win you might bet against it. This will change the market and Hilary might now only have a 89% implied probability of winning. This will continue up to the point where the whole market information is aggregated, and more accurate percentages will arise. The other case is the replicability of academic results. Asking experts whether the results of a paper can be replicated shows worse results than letting the public bet on in with a prediction market. Again, it incentivizes people to share their knowledge, by earning money, showing better results than the opinion of experts. Experts might be biased, and don’t want to offend colleagues, while prices and profits never lie.

We have now discussed that price markets for classic financial goods conveys vital information. We have also discussed that prediction markets can incentivize the public to reveal their knowledge. But as you might have guessed: I am not particularly interested in those markets. The things that interest me are marriage markets, markets for pricing human lives, and so on. So why not put a pricing mechanism on those things to better predict the future.

About 45% of marriages end in divorce. As an individual, the decision to marry someone is one of the biggest investments imaginable, emotionally, and financially (See Marriage as a Commitment Device or Beauty and the Social Market). I would put a percentage on how likely it is for my marriage to last. But my partner also has her percentage. My parents, her parents, our friends, and even outsiders might have an opinion on whether this marriage is doomed. I would like to know these opinions. One thing is clear, I won’t get real answers when asking them. If I’d have to reveal my likelihood of divorce to my future wife, I will most definitely lie. Instead, I would create an anonymized prediction market, where people can buy and sell futures contracts on when they think this marriage will end. This will incentivize everyone including the about-to-be-wedded to reveal their true beliefs. This market would aggregate all the knowledge available. Finally, I can update my opinion on whether this marriage will work out. I can then take measures to either invest more heavily or hedge for divorce more stringently. The information has been pooled eloquently and people will win and lose money. Everyone can update their prediction of the future, thus, live a more efficient life (YAY).

Same thing for peoples’ lives. I would love to have a prediction market on when a person will die. Again, there would be a contract-based futures market on the timing of death. First of all, this would show the best possible pooled knowledge of when this person will die – so I can update my expectations on how to treat this person and how much to invest in this person. Furthermore, I could hedge against an early death of vital co-workers, friends, family, and loved ones. I could bet on their early deaths and maybe earn some condolence money as compensation for their loss. In absurd cases, one could bet on a late death of enemies or politicians one hates. This way, if they live a long life annoying you, at least you get money out of it.

The biggest problem to date with prediction markets is fraudulent behavior. People can cheat. In the two cases I mentioned: A married couple could bet heavily on divorce and then get divorced as a result. In the case of prediction markets on life, this could be misconstrued as putting money on someone’s head or hiring a hitman. In the end, you are giving another person money via the contract if that person dies. To be honest I am not too concerned about this. I think people will realize which markets can be fraudulent and it will be prized in. I don’t mind insider trading, or cheating, as long as everyone knows that this could be subject to just that. This social prediction market could open a whole new world of efficiency. But unfortunately, it might just stay a wet dream of mine.