The deficit and empirical game-theory research

I've had my three-hour meeting for the day (yes, on Sunday), so I'm rewarding myself with a small square of spicy chocolate and a blog entry on politics. The Commission for Very Serious Person Theater is missing a huge chunk of government spending by focusing only on the federal budget deficit and ignoring state spending patterns (I am quite sure that someone like Matthew Yglesias or Dana Goldstein has already pointed this out), but it's an important issue in itself unless you want health care costs to swallow up all of federal spending in a few decades and (almost as important to an academic) an interesting problem in game theory: when the political incentives are in favor of posturing (and that means either hyperpartisan posturing or Very Serious Person theater), what mechanisms might create some realistic policy?

As a union activist, I have a reason to occasionally check on research in areas such as game theory and conflict resolution because, well, collective bargaining is a political process with some areas of mutual benefit and some areas of zero-sum dynamics either in reality or in the political sense. Add in inevitable miscommunication (either minor or major), plus Florida law requiring all bargaining be conducted in public, and you have a recipe for one of the most inefficient interparty negotiation structures ever.  So I have a Swiss-cheese-like knowledge of some practical approaches to conflict (i.e., what's distinctive is the number of large holes). 

And it strikes me that given the unreality of current budget discussions, a tool such as this weekend's New York Times interactive "you close the long-term deficit!" game promises one foundation to addressing the issue or at least conducting some interesting empirical research in game theory. Matthew Yglesias's choices use changes to retirement-benefit ages in federal policy (though he says such changes are about the most regressive option possible). Mine do not. We're both liberal, and we both take the gains about equally from tax increases and spending cuts. I suspect we could come to some sort of agreement quickly.

But Paul Krugman and Grover Norquist? Hmmn… Time at least in theory for the Brams-Taylor procedure. It's patented, so you'd have to call upon Fair Outcomes (the operator of the patented procedure), which charges very fair rates, certainly a tiny fraction of any commission to follow up on the Commission for Very Serious Person Theater. It's a form of a bidding procedure: in a two-party case, the parties divide up a complicated issue into a set of discrete pots, define what "winning" each pot would be, what dividing a pot proportionately would be, and essentially bid on winning the various pots (by allocating 100 points among the pots in a sealed-bid format). The algorithm decides who wins pots and then how some pots are divided in a way to maximize each party's gains as defined by their bids. 

I strongly suspect that while Krugman would have few problems agreeing to participate in such an exercise, few involved in politics would commit in advance to supporting a procedure that would end up undermining a purist advocacy position such as Norquist's drown-the-baby approach to government. But that doesn't prevent academics interested in the exercise from conducting it. (No, not me: it's far from my area of research, and I'm overcommitted.)

Two potential flaws in the approach: (1) there are some ways to game the system if you know the other party's preferences and decide to misrepresent your own for tactical benefit; (2) it is highly plausible that an approach that maximizes a very conservative person's goals and a very liberal person's goals on specific spending and taxing issues will not end up balancing the deficit by 2030 or coming at all close. The first problem is solved by asking people whose jobs are not to win elected office or engage in Very Serious Person Theater to participate in the exercise (i.e., participants in a research study whose identities are never disclosed). The second is easily remediated: have "closing the budget deficit" be one of the goals to include it in the maximized items, and if neither party in each trial sees it as important… well, we've learned something, haven't we?