Friday, October 12, 2012

Meanwhile, Back at the Cliff...



Well, I had fun watching the VP debate last night. But let's say--just for the sake of discussion--that the Bluegrass Rumble did not fundamentally transform the electoral landscape. Where does that leave us? Specifically, where does that leave the prospect of a major deal on the bundle of fiscal decisions known as the "fiscal cliff"?

Yesterday (about noon) I took the election odds from Intrade for the White House, control of the Senate, and control of the House (see details below). For the Democrats, this yields these % chances of Democratic success (Republican odds are the remainder):


Obama win 63.3%
Dems control Senate 69.5%
Dems win House 11.6%

The striking implication is that the 2012 elections are highly unlikely to grant either party unified power to implement its preferred policy. There is about a 5% chance of Democratic control, 10% chance of GOP control.

There is about a 39% chance that the 113th Congress will repeat the current political alignment, and 46% chance there will be some other form of divided government. This last estimate is probably inflated since I assume these events are independent, whereas in real life a GOP surge to win the White House would probably be accompanied by increased GOP success in House and Senate races (for example). 

I just want to make two quick points about this estimate.

1) as a citizen I would be really disappointed if the current power alignment continues. The 112th Congress has been abysmal. The Worst. Congress. Ever. During a time when the U.S. has faced immense challenges, the Congress has been essentially paralyzed on major policy problems. Someone ought to pay a retrospective price for the failure. Maybe voters want to blame the Republicans, maybe they want to blame the Democrats, but one party or the other ought to pay for this failure. Prospectively, a continuance of divided government threatens to mean more of the same. If the same players are reelected, that's a mandate for more of the same.

2) If elections won't break the partisan logjam, Congress should think about Plan B: compromise. When I first took a game theory class, I was relieved to learn that when it is costly for actors to procrastinate, they will avoid those costs by cutting a deal as quickly as possible. Hah! There are more sophisticated models that explain why we actually observe wars, strikes, and gridlock, but the 112th Congress could learn a bit from the simple models. If elections aren't going to change the identity or preferences of the main players, there's no point in procrastination. If Rs & Ds just struck the deal they have been flirting with for two years--a blend of tax increases and spending cuts--it would cut the unemployment rate by .5%, which is .5% better than voting to repeal Obamacare again

There doesn't seem to be any Congressional effort to work on a fiscal cliff bargain. After a five-week August break, Congress returned, agreed to punt the fiscal year 2013 appropriations bills into the 113rd Congress, and got back onto the campaign trail. Nor it is likely that there will be a lame-duck deal because Speaker Boehner is waiting for, well, nothing.

Details
* I used the likelihood of Democratic victories and, instead of using the markets for Republican victories (since total odds can sum to more than 100%), I assumed the Republican odds for the White House and House were (100% - Democratic odds).  For the Senate, there was a 15% chance of a tie. In the event of a tie, the VP decides the majority control, so I divided up this "tie" prediction between the two parties based on the White House prediction: +9.5% for the Democratic % chance to control the Senate, +5.5% for the GOP.

6 comments:

  1. Game theory only predicts that parties will compromise if both parties feel the cost of compromise will be less than the cost of not compromising. But for the Republicans in Congress, there seems to be a distinct belief that any compromise whatsoever will be worse for them (individually and collectively) than suffering whatever disaster results from not compromising. (Worse for the country, of course, but that is clearly not a consideration for most of them.)

    Which means that the odds of disaster are rather higher than your calculations would suggest.

    ReplyDelete
  2. These calculations are exceedingly sloppy, based on a plainly faulty assumption - that each party's success in the presidential elections, Senate elections, and House elections are independent events. That's so obviously not the case. The precise nature of the relationship between those events is quite complex, and could probably be best estimated using some kind of model based on historical data and analysis. But it's absurd to assume there's no relationship. Your calculations clearly underestimate the likelihood of the "unified Dem" or "unified GOP" outcomes (based on the intrade numbers that you use). For example, if the chance of the Democrats winning control of the House is 11.6%, most of the scenarios under which this would happen would fit within a significant national shift in favor of the Democrats, in which they would be nearly assured of winning the Senate and much more likely than 63.3% of winning the presidency. I'd guess that the probability of your "Unified Dem" scenario would be somewhere around 10%, as opposed to your 5%, although the proper way to estimate it would be to run some kind of simulation of the kind that Nate Silver does (I don't think that he currently does any that would give these estimates, but his models could be adapted/augmented to do so).

    ReplyDelete
    Replies
    1. Steve, I am going to mark you down as agreeing with this passage: "There is about a 39% chance that the 113th Congress will repeat the current political alignment, and 46% chance there will be some other form of divided government. This last estimate is probably inflated since I assume these events are independent, whereas in real life a GOP surge to win the White House would probably be accompanied by increased GOP success in House and Senate races (for example)." Yes, a more sophisticated model might predict somewhat higher probabilities of unified government, but the main point holds: it currently appears highly unlikely that either party will be able to resolve the budget impasse without striking a grand bargain with the opposing party.

      Delete
  3. I don't quite agree with the numbers in that statement, but I do agree that those two cases are more likely than the unified cases. Personally, I'd guess (based on the individual Intrade #s you use) the likelihood of the "Unified Dem" scenario to be around 10% and of "Unified GOP" to be around 20%, which would leave 70% for the other two (as opposed to the 85% that yours sum to). I do agree with the general points & observations that you make, my only issue is with the numbers in your pie chart, and the fact that you omitted your most important (which happens in this case to be unrealistic) assumption - that these events are independent - from your "details" note explaining your calculations.

    ReplyDelete
  4. The prospect of a major deal on the bundle of fiscal decisions known as the "fiscal cliff"?

    debt collection agency

    ReplyDelete
  5. I concur with your conclusions and will thirstily look forward to your upcoming updates.


    hanging pot racks for kitchen

    ReplyDelete