If (On Court Resignations, Nominees, and Information Markets)

If there is no serious misstep in the next couple of weeks, Samuel Alito will be confirmed as a Supreme Court Justice, and Sandra O'Connor's iffy resignation will become real. I think such "ifs" are provocative.  First, we should wonder about the future of contingent resignations.  O'Connor is sitting on the Court now, and this sort of lame duck could encourage strategic behavior (both in confirming a new Justice and in deciding cases on the Court), though a straightforward resignation, and an 8-member Court, might simply have raised other strategic possibilities.  Leaving aside the idea that the Court could have 10 or some other number of members, the contingent resignation might next be taken one step further: a Justice could resign subject to the confirmation of a successor by a certain date.  Thus, a conservative Justice might offer to resign, contingent on the sitting Republican President appointing someone and seeing that nominee confirmed.  We might squirm if we observed such a conditional offer of resignation, but it is only one step beyond where we are now.  Another would be a resignation contingent on seeing (and approving) the identity of the next nominee.  That seems almost beyond imagination.

But would it be beyond imagining for the Senate?  If Senators are inclined to be strategic, then we should expect some no votes where there is an expectation that the next nominee would be more to the voters' liking.  And we should expect yes votes where the current nominee is perceived as more desirable than the likely replacement nominee.  In turn there is the question of how Senators can gain information about likely subsequent nominees.  Tradesports.com tells us that there is a 78% chance that Alito will be confirmed, but it does not yet provide information as to who is likely to be nominated if Alito is not confirmed.  It is difficult to develop markets about such contingent information, for they will be thin.  In principle, there could have been a market in the likelihood of Alito or McConnell or Gonzalez being confirmed even as Miers was making the rounds in the Senate.  One market would predict Miers's likelihood of confirmation; another would predict Alito's, and so forth.  Putting these markets together would, for better or worse, likely change the actual votes on the first nomination - which would in turn be reflected in the market prices.

Some markets are thick enough that no contingent markets are desired.  For example, there is a market in oil and it reflects the likelihood of hurricanes and wars and so forth.  There is also (now) a developing information market in such things as the likelihood of a war or a hurricane itself.  If I want to aggregate dispersed wisdom and information as to the likely price of oil in the event of a particular war, I can reason from the behavior of these two markets.  But where there is no thick market, I might need, and pay for, an information market that focused on a particular contingency.  Budding information markets may the best predictors of the next Republican nominee for President, and they might also be good predictors of the likelhood that more than 100,000 U.S. troops will be in Iraq in November 2008.  But we might expect a separate market on the question of who is the most likely nominee in the event that there are (or are not) that many troops in Iraq. It will be interesting to see whether such contingent information markets flourish, much as it will interesting to see the evolution of contingent Supreme Court resignations.

Comments

[The Law Fairy]

The idea of a market forecasting contingent nominations and similar political happenings is both interesting and troubling. How might such a market have a, if you will, Heisenberg-esque effect on the very events it predicts? Much like stock market tamperers (as we've seen recently with Martha Stewart and the like) can through calculated mis-information *cause* the events they predict, might not a market like the one you hypothesize in effect create the very news it seeks to predict? One could, of course, make the argument that such things already occur -- I was particularly intrigued by Professor Sunstein's recent post about information cascades.

Even if the market could somehow be regulated to try to prevent tampering, couldn't even legitimate market participants affect the course of events? If so, is this something we would want to encourage or discourage? I'm wary, particularly keeping in mind that those with the greatest ability to dominate such markets tend also to have vested interests in the outcomes they research.

[Eh Nonymous]

I was fascinated to see SDO'C's contingent resignation, as it brought home an important point:

While the method appointment of Justices is constitutionally determined, and similarly of their (involuntary) removal, the mechanics of their resignation and replacement are less fixed.

There are rules - Supreme Court rules? Federal rules? - about when a Justice is given that spare office, how many clerks they get, how much pay and for how long, and when the other perks of office do and don't apply.

I could imagine a highly choreographed contingent resignation, involving the stock market, the identity of the person in office, the identity of the replacement nominee, the house the Moon is in, and whether a majority of the sitting Court approves. Is it unconstitutional? Well, the Senate and President have no power to fill a vacancy that doesn't exist, and the sitting Justice has all the power to create that vacancy if they want.

So, why not?

[FXKLM]

The interesting thing about a conditional resignation with a time limit is that it completely changes the incentives for obstructing nominees in the Senate. With O'Connor's conditional resignation, Democrats have very little incentive to vote against reasonably moderate nominees. Even if they manage to block one nominee, the next one will be chosen by Bush also. Since Democrats can't seriously hope to delay the process until 2009, the benefits of obstruction are limited.

But imagine that Justice Stevens offers to resign contingent on a successor being named within two months. Democrats would probably be able to filibuster until the deadline expired without paying too high a political price. If Bush couldn't get someone confirmed in time, he would lose the opportunity to replace Stevens, but Democrats would run the risk that Stevens would either die or retire without such a contingency under a Republican administration.

I think that's probably a more plausible scenario than setting a deadline that coincides with a presidential election. It's less obviously political, and the sitting justice could plausibly claim that he or she wants to avoid a long, ugly confirmation battle. I'm not sure how it would play out, but it would definitely be interesting.

[MMJ]

Given that Supreme Court nominations, Supreme Court confirmations, and Supreme Court deliberations are governed by strategies to advance political ideology, it shouldn't seem so unusual that Supreme Court resignations could also be governed by such considerations.

But, I think there's a huge difference between resignation contingent on [the Senate's] confirmation of a successor, and resignation contingent on [the Justice's] approval of a successor. Resignation contingent on approval would, in effect, create an added step in the selection process, one I'm guessing is not mentioned in the Constitution.

It might have to be decided by the Supreme Court. I honestly wonder how they'd vote.

[Jason Ruspini]

Law Fairy, You bring four separate questions to mind:

1) What is the influence of manipulators on price?

See Hanson and Oprea's "Manipulators Increase Information Market Accuracy" http://hanson.gmu.edu/biashelp.pdf , one of the most interesting papers on prediction markets. They use a micromarket model to show that a manipulator's price bias, if known, has little effect on market prices, but variance in the manipulator's bias can actually increase accuracy, by enticing more liquidity. As with all models, there are many assumptions here and the ones I note relate to the sizes of various traders' budgets and their risk-appetites. What about cases where there is a natural imbalance due to external factors? If some traders are risk-averse, they may make trades in a specific market for reasons other than their best estimate of price. If they have relatively large budgets, this could go on for some time.

2) What is the influence of insiders on both price and outcome?

A problem in any market, but much more so in markets such as those linked to terrorist-related events, *where anyone could become an "insider."* In those cases the settlement process would have to be especially rigorous, and this would be advertised beforehand. In cases where outcomes are determined by a small pre-defined group, the insiders can more easily be excluded from the market. (Prediction markets can be unreliable in these cases, especially when there is a semi-open set of possible outcomes, e.g. a list of candidates for some office. -- and note the Alito leak.)

3) What is the influence of price on price?

How prevalent is feedback and momentum trading? Is this "noise" normally distributed? How to distinguish serial correlation caused by such trading with moves caused by sustained changes in "fundamentals"?

4) What is the influence of price on actual outcome?

Why I placed "fundamentals" in scare quotes above..
This is the most difficult and interesting of the questions. Nash inflation comes to mind, and of course, Soros' theory of reflexivity. *When you separate this question from #2 above, it doesn't seem as scary though, does it?*