Bryan Caplan  

Betting Rules

PRINT
Does Paul Krugman believe that... End Draft Registration...
Tyler on the meta-lesson of our unemployment bet:
I think this episode is a good example of what is wrong with betting on ideas.  Betting tends to lock people into positions, gets them rooting for one outcome over another, it makes the denouement of the bet about the relative status of the people in question, and it produces a celebratory mindset in the victor.  That lowers the quality of dialogue and also introspection, just as political campaigns lower the quality of various ideas -- too much emphasis on the candidates and the competition.  Bryan, in his post, reaffirms his core intuition that labor markets usually return to normal pretty quickly, at least in the United States.  But if you scrutinize the above diagram, as well as the lackluster wage data, that is exactly the premise he should be questioning.
Let's analyze this point-by-point.
I think this episode is a good example of what is wrong with betting on ideas. 
I think this episode is a good example of what is right with betting on ideas. 
Betting tends to lock people into positions,
Tyler's not wrong.  Betting "locks people into positions" in the laudable sense that it makes them clearly state what their positions are.  Once positions have been spelled out, though, betting spurs the opposite of dogmatism.  The clarity of the resolution fortifies right people and disconcerts wrong people.  That's progress toward truth.
...gets them rooting for one outcome over another,
Indeed.  But unless either party has the power to change the outcome for the worse, what difference does it make what they're rooting for?
...it makes the denouement of the bet about the relative status of the people in question,
Partly, but this is all for the good.  The loser had a theory with a prediction he was willing to bet on.  He was wrong.  This isn't conclusive proof his theory was wrong or his judgment is poor, but it is an ironclad reason to reduce confidence in the theory and the judgment of the person who held it.
...and it produces a celebratory mindset in the victor.
I fail to see why this is a problem.  The possession of demonstrable insight is worth celebrating.  But if victory laps are bad, they are a far lesser evil than what pervades the betless badlands: the eternal celebratory mindset of pundits who refuse to specify in advance what counts as victory.
That lowers the quality of dialogue and also introspection, just as political campaigns lower the quality of various ideas -- too much emphasis on the candidates and the competition.
There's a world of difference between elections and bets.  Elections are popularity contests.  Bets are cognitive contests.  The electoral mistake is to conflate the two: To think ideas are right because they're popular.  I've never known a victorious bettor to make the parallel mistake - to think that because they've won a bet, the political system will heed their wisdom.  I personally expect nothing of the sort.
Bryan, in his post, reaffirms his core intuition that labor markets usually return to normal pretty quickly, at least in the United States.  But if you scrutinize the above diagram, as well as the lackluster wage data, that is exactly the premise he should be questioning.
I'm happy to admit there are broader questions here, though Noah Smith, not Tyler, has the right of it.  But when a bet lets us make progress on narrower questions, we shouldn't rush to change the subject.  Instead, we should pause, reflect, and be grateful that we genuinely learned something.  Betting rules.




COMMENTS (19 to date)
Rick Hull writes:

FWIW, I agree with your points here, but your celebration laps in the prior post struck me as a bit tone deaf.

RJ Miller writes:

I have to agree with all your points, especially the question of whether or not betting "locks" people into anything.

Why would anyone choose to pigeonhole into a viewpoint they don't entirely agree with rather than add nuance or not bet at all?

James writes:

"Why would anyone choose to pigeonhole into a viewpoint they don't entirely agree with rather than add nuance or not bet at all?"

Because they identify a positive expected return. If there is nuance, they can adjust the size of the bet or diversify across bets. I don't "entirely agree" that any of my investments will perform well but still I hold them. You probably do the same.

Michael Savage writes:

Accepting 20:1 odds on an outcome I think is 10% likely is a good bet, but one I still expect to lose. Losing doesn't make me wrong; it was the outcome I expected. I know you know this, but the tone of your comments seems too absolute to me. Winning a single bet doesn't necessarily demonstrate insight and shouldn't necessarily increase status.

I'm mostly with you on betting, and I've become very unpopular for challenging people to put their money where there mouths are, but I have sympathy with Tyler's views here. Some questions are uncertain rather than probabilistic, and don't easily lend themselves to betting. It favours simple heuristics like 'reversion to mean', which may be a sensible prior, but which militates against the kind of speculative thought that sometimes produces really new insights.

Steve Brecher writes:

If by "lock people into positions" Tyler meant what Bryan infers -- clearly state positions -- then that's true, but trivially so. But if Tyler meant rigidity of position going forward, as implied by the associated comment about "rooting for," then it's not (necessarily) true. A bettor who changes or softens a position can lay off (part of) the bet with another counterparty, or offer to buy it back at a negotiated price with the original counterparty.

Sieben writes:

What Tyler means is that betting locks you into a position for a long period of time. Presumably this is bad because you shouldn't feel free to change your mind at will. But revising your opinion once you have stated it clearly and put money on it probably messes with your (self)image as a consistent person.

Philo writes:

Caplan 1, Cowen 0. (Take another victory lap!)

Pat writes:

Cowen usually avoids clear statements of his positions. I'm surprised he ever took a bet with you.

Pat writes:

I also don't understand his defensiveness. His position was he thought the chance of his position being correct was greater than 9%. If I expect to lose a bet 9 times out of 10, I would just shrug and not protest too much.

Michael Stack writes:

The whole thing is unseemlly. I'm fine with betting, but Bryan's gloating (in his previous post, not this one) seems to be contrary to his whole "spirit of betting" that he's championed.

Similarly, I would have put more stock in Tyler's complaints about betting had they occurred prior to his loss. In this context, it sounds like sour grapes.

Nacim writes:

I don't understand the accusation of gloating on the part of Caplan, he explicitly commends Cowen for willingly taking a bet. Cowen just strikes me as annoyed that he lost, especially when he complains about how "it makes the denouement of the bet about the relative status of the people in question, and it produces a celebratory mindset in the victor." I also fail to see the issue with why that's a problem. Good job on being correct Bryan! Hopefully you encourage greater insight and accuracy in others.

Brad writes:

@Pat

I believe it was Tyler that gave 10-1 odds. He thought that it was a 10-1 bet that unemployment would not go under 5% at any point over the next 20 years.

He in fact VERY strongly believed he would win the bet most of the time.

Michael Stack writes:

@Nacim,

You may have a point. I went back and re-read Caplan's original post and it didn't seem as gloating as I'd originally remembered. My apologies to Caplan.

ThaomasH writes:

What is generally wrong with most bets is that they are not conditional. A and B may disagree on the conditional response of the economy to Feb policy X and differ about whether the Fed will in fact adopt policy X. The outcome of a bet in those circumstances may not be very enlightening. Did A win because his economic model was better or because his policy model was better? Of course no bet can specify all of its conditions, they would be better if they tried.

Nathan writes:

Cowen's argument now seems to be that Caplan was right in the specifics but wrong in general - that there is a genuine phenomenon occurring that Cowen expected to be reflected in unemployment numbers but is in fact being reflected elsewhere.

Instead of sour grapes-ing Cowen should propose a modified bet based on his refined theory.

Philo writes:

He should propose a modified bet; but he won't, because he's down on betting. (He says he doesn't like the "celebratory mindset" produced in the winner, but it seems he especially dislikes the *abashed* mindset of the *loser*.)

Pat writes:

Brad,

Thanks for correcting me. I was way off!

Pat

Mark V Anderson writes:

I agree with Bryan that betting can help clarify a position by making it precise enough to make a bet.

But I agree with Tyler that it does tend to lock in the betters to their positions. It is already difficult for people to change their minds on views that they've stated publicly. Betting on the position makes it even harder to change because it makes your position a matter of winning or losing more than being right or wrong.

Garrett M writes:

[Comment removed. Please consult our comment policies and check your email for explanation.--Econlib Ed.]

Comments for this entry have been closed
Return to top