What went Wrong with Healthcare?

No, I’m not talking about the de/merits of the substantive decision on healthcare, or our overall healthcare system, both of which I am only marginally qualified to discuss.*  No, rather, I’m concerned with why so many predictions about the decision were incorrect.  I’m not talking about individual media pundits, or people who wrote blog posts with predictions.  Rather, I’m talking about the erroneous decisions of the prediction markets.  In the days leading up to the decision, I was checking the wisdom of the crowd on both Josh Blackman’s Fantasy Supreme Court and Intrade (for the record, Intrade was much more wrong (75% chance that the law was unconstitutional), but both were wrong). 

I’ve written before about prediction markets and the Supreme Court, and how I think these markets might provide us with a valuable service.    There’s a value in having predictable outcomes to cases.  Holmes, I think, would support my view that predictability is akin to the rule of law.  So why did this amazing new technology not work properly in this instance? 

I don’t have a clear cut answer – yet – but I think some of it has to do with the way the prediction contracts were phrased, the oral argument, and the difficulty of predicting one person’s actions.  First, prediction markets are crude tools that give us a “yes” or “no,” and not very good at trying to get at the reasons behind a ruling.   Second, the oral argument was a turning point for many participants in the markets, even though there has been some discussion in the theoretical literature indicating that oral argument has little, if any effect, on actual outcomes.   So it could be that participants were misled by a factor that was illusory, and once people started saying that the law might be overruled, it touched off an irrational feeding frenzy.  Another possibility would be that prediction markets do not do particularly well with decisions that are left up to one person.  After all, there is such a thing as free will.  Just because we have prediction markets, it doesn’t mean that we can read someone’s mind.  That’s why the prediction markets failed so badly to predict W’s nomination of Harriet Miers to the Supreme Court.  It’s difficult to get into the mind of the president or his administration.   Without inside information, the crowd is basically all equally uninformed.   And so it was difficult – very hard for the crowd to predict – Robert’s decisions or reasons for voting as he did.  I’ll be thinking more in the coming days about how prediction markets could avoid the trap they fell into with this decision.

*I say only marginally qualified because I teach contracts, business associations, and employment law, with a twist of all things tech and global, not constitutional law or health law.   I have, however, read the decision, which puts me head and shoulders over a large number of television commentators yesterday  (“I haven’t read the decision yet, but assuming that the court said X…”).  Finally, speaking of “getting it wrong,” this post would not be complete without a big ole turkey to Fox and CNN for reporting the holding incorrectly.

You may also like...

12 Responses

  1. Howard Wasserman says:

    Maybe it’s a version of what Justice Ginsburg said a few weeks ago: Those who who are betting don’t know and those who know aren’t betting

  2. PrometheeFeu says:

    Here is an alternative: the markets weren’t wrong. Intrade correctly evaluated a 75% ex-ante probability of the mandate being struck down and this case turned out to be in the remaining 25%. The market wasn’t wrong, just unlucky. Prediction markets don’t actually predict the future, they just aggregate knowledge.

    Consider a prediction market on the roll of a die. The market would be correct to price itself at an 83% chance of the result not being a 6. That market does not become wrong just because the die comes up as a 6.

    Of course, this is unappealing and I think inaccurate. (thought the reasoning holds in general) I think a 75% probability was way too high. My guess is that participation in prediction markets is correlated with a certain preference for free-markets and libertarianism. Such people would be more likely to want to believe the mandate unconstitutional and be ready to pay money to demonstrate that fact.

  3. A.J. Sutter says:

    If we accept PrometheeFeu’s point, which I think is on the right track, then what may be misleading some people is to call it a “prediction market,” rather than a market estimating probability distributions. Moreover, even though it aggregates “knowledge,” it’s in a condition of extreme knowledge asymmetry, as Justice Ginsburg mentions. So it’s not particularly high-quality knowledge. Moreover bis, unlike a die, which can be rolled under more or less repeatable conditions, a ruling such as this is not repeatable. So from a frequentist point of view, what the “market” is coming up with isn’t even a probability distribution: it’s simply an opinion distribution, under conditions of extreme information asymmetry.

  4. Brett Bellmore says:

    Another point is, you might correctly estimate the odds in, say, a boxing match, based on past performance. You’re still going to get blindsided if the champ decides to take a dive.

    The prediction markets were right, it turns out: Roberts had the votes to win. He decided to lose, instead. He took a dive.

    Who can predict that sort of thing? Everybody thought it hinged on what Kennedy had for breakfast, not whether the Chief Justice would decide to twist himself into a pretzel finding an excuse to uphold a law he surely thought was unconstitutional.

  5. Joe says:

    Back in the 1790s, though I have no idea if Brett ever thinks the Constitution was actually correctly applied,* justices noted that they would only overturn federal laws when they were absolutely sure of their position. Roberts didn’t feel he met such a heightened test that factors in to how justices carry out the weighty activity of overturning federal law.

    * For those not aware, Brett has alluded to his belief that currently the courts simply don’t properly interpret the Constitution, at times suggesting some time (like the New Deal) was the turning point. Randy Barnett et. al. also think the ND was a nadir. Not seeing the days beforehand as ideal, I find this sentiment dubious.

  6. Actually Josh Blackman’s platform had it about right – lots of uncertainty gets you the ~54% to ~46% split that was present before the decision.

    A useful question to be asking here is why the compositional differences in the respective crowds (Intrade v. FantasySCOTUS) generated the spread between {27%=Affirm Intrade} vs. {46%=Affirm FantasySCOTUS} … and turn how that might be leveraged to generate a better prediction engine.


  7. Shag from Brookline says:

    Brett should be aware that the decision making process at SCOTUS is not as simple as ordering a mail-order-bride. Brett accuses CJ Roberts of taking a dive. But he fails to accept that Justices Scalia, Thomas and Alito have long been in the tank for conservative causes. No, CJ Roberts, did not take a dive, he merely refused to Join the Etna cesspool. And Justice Kennedy apparently doesn’t like competition in swinging. Yes, Brett continues with his chronic case of “Wick-burn,” for which there is no soothing balm.

  8. Miriam A. Cherry says:

    Thanks, everyone for some interesting thoughts/comments. Yes, the Josh Blackmun / Fantasy SCOTUS market was more correct. He seemed to think that it was because of the way that the questions were phrased on that market versus the way that Intrade phrased it. Very interesting!

  9. Another possibility would be that prediction markets do not do particularly well with decisions that are left up to one person. After all, there is such a thing as free will.

  10. Buena información, valioso y excelente diseño, como compartir cosas buenas con buenas ideas y conceptos, lotes de gran información e inspiración, que necesito, gracias a ofrecer una útil información aquí.

  11. Metler Law says:

    “the crowd is basically all equally uninformed” – great quote. I feel this way about the majority of politics today.

  12. Thank you for the great post,It is really a big help.thanks for sharing nice blog .