Bryan Caplan  

Why Not Raise the Stakes?

The Park Ranger Metaphor, Furt... Thoughts on Charity...
In the comments, rapscallion poses an interesting challenge:
The bets I've seen placed on this site are so small in value that they don't tell me much other than that none of the bloggers believe in anything with much certainty. A couple hundred dollars is nothing to upper-middle class academics; I'll know you're seriously convinced of something when you start placing bets of $25k or more--enough money to really hurt.
I disagree.  If a couple hundred dollars were "nothing" to upper-middle class academics, getting them to bet would be easy.  In practice, it's like pulling teeth.  See how hard it was to find anyone to bet me $100 on "the bailouts will lose money" or "the U.S. will have lower unemployment than Europe over the next ten years."  The main reason I don't raise the stakes is because I want people to actually bet me.

But why are small sums enough to deter 95%+ of the people who disagree with me?  I see two main reasons:

1. We aren't just betting $100.  We're betting $100 plus reputation plus bragging rights.  That's why I prefer to bet the famous.  The Simon-Ehrlich victory wouldn't have been nearly as awesome if Simon bet a random Malthusian.

2. Many spouses, perhaps most, disapprove of betting.  They think you're irresponsible when you bet, and stupid when you lose.  Imagine how badly they'd react if the stakes were $25k!  Even the victor might find himself stuck in the doghouse.

Comments and Sharing

COMMENTS (13 to date)
Zac Gochenour writes:

I agree with rapscallion.

Imagine if the bet were $1. Many people would still say no simply because they are unwilling to bet any amount of money; they just aren't interested in betting. At your income, infrequent bets of $100 isn't much different from $1.

When you bet a very small sum like this, all you're really saying is that you're the kind of guy who likes betting -- that you think it is fun. It doesn't signal the intensity of your belief or make you "put your money where your mouth is," -- its no different from betting only on reputation and bragging rights. The idea behind adding the monetary portion is that it has to add to the signal. $100 just doesn't cut it.

I think you should find people who are making extraordinary claims and then offer them much better than 1:1 odds. For instance, your bet with Arnold was very much in your favor: as I pointed out at the time, you could have hedged your bet with intrade and made a profit no matter the outcome. Imagine if you offered him 5:1 odds and raised the bet to $1000. I think that's really signalling something.

razib writes:

In practice, it's like pulling teeth.

that's been my experience too. i like the idea of betting, and straight up ask people what odds are on all sorts of topics. i generally like to bet against things i'd like to succeed, which is a win-win (and lose-lose) situation for me. but it is really hard to get people to put money down. usually i bring money in when people express strident certainty about something.

Tracy W writes:

At your income, infrequent bets of $100 isn't much different from $1.

Huh? It's $99 more. How much do you think that professors earn? Bryan Caplan's books are not up there in the sales rankings with J.K. Rowlings, which may be the fault of society, but still doesn't mean that he has the sort of income where there's little difference between $1 and $100.

8 writes:

He must not have seen Trading Places.

Zac Gochenour writes:

@Tracy: I can basically guarantee $100 is less than half a day's pay. It is the sort of money that might be spent on a nice dinner for the family.

The purpose of the monetary bet is to get people to put their money where their mouth is - to get them to risk something more than just their reputation. If you're an academic making a bet on some claim like this, I'll go out on a limb and say your reputation is worth considerably more than half a day's pay.

Just think about this in Bayesian terms. Say Caplan makes some prediction and challenges someone else of prominence to bet him for bragging rights. Assign some probability to him really believing his claim. Now update your probability given the bet also includes $1 at even odds. Now update it with the bet being $100 at even odds. Are the last two probabilities significantly higher?

Thinking about this more, I think that at best I'm no more confident after the bets, but the wager amounts are so small I may be less confident. If you're really confident, why not bet more? If the answer is, "because my opponent wouldn't take the bet," why not offer more attractive odds and then bet more?

If you are just betting reputation, your underlying level of certainty is unknown. When you bet a small amount at even odds, you could be specifying "I think this is about 50% likely," or even less, if you're risk-loving at that amount.

Tracy W writes:

I can basically guarantee $100 is less than half a day's pay. It is the sort of money that might be spent on a nice dinner for the family.

To me, that sort of money is money not to be sneered at. If something is the pay for 5 minutes work, that's not worth worrying about. Half a day is significant.
And I note that you'd be hard pressed to get a nice dinner out for a family for $1. While we're at it, you'd be hard pressed to get a nice dinner in for $1.

What sort of lifestyle do you lead, where you'd spend half a day's pay as blithely as a few minutes pay?

The rest of your comment is based on this idea that there isn't any real difference between $100 and $1, as I don't share your starting point my opinions are the opposite.

David W writes:

Zac - it seems to me that there's an easy solution for your quandrary. Just find a topic on which Bryan is extremely confident and you believe he's wrong, and offer the bet yourself.

I mean, the whole point of this discussion is that Bryan's willing to put his money where his mouth is; can you say less?

rapscallion writes:

I agree with you that reputation and bragging rights can be significant incentives. Since these things are usually pretty important to academics, I’ll concede that just the fact that a professor is willing to make a public bet usually indicates that they think a proposition is more than, say, 60% likely to be true.

However, I think there are some important rejoinders:

i) Your point 1 doesn’t really explain the low-stakes monetary values; it just argues that subjective p-values can be high despite low-stakes. Just because profs care a lot about reputation, it doesn’t mean that they don’t care about money. Really, your main explanation for the lack of high-stakes bets is simply that spouses don’t allow them.

ii) I doubt that the number of potential bettors against a proposition is always decreasing in betting value. When you start getting into higher dollar amounts, you might lose the attention of all the esteemed professors, but people outside academia, like professional gamblers, might take notice. Yes, winning bets against these people might earn you lower reputational gain than against esteemed profs, but money is still nice despite the source, right? You might also be able to make money off kooks; the skeptic James Randi has gotten a lot of psychics to try for his million dollar prove-you-have-magical-powers prize.

iii) I still think that if beliefs really matched rhetoric, high-stakes bets would be more common in spite of the additional costs imposed by spouses. Would you really sweat a bet that the sun would rise tomorrow? That the germ theory of disease is true?

iv) Are single/divorced profs more likely to make bets?--Not as far as I can tell.

eccdogg writes:

To me the key to betting is not so much the dollar amount but the contract.It forces people to explicitly specify what they think will happen, when and with what probability. It really focuses the debate about issues and provides a clear winner.

Also it is almost entirely about reputation.

I work on a trading floor and not supprisingly we bet about a lot of things. And the bets are often as small as a quarter and this is for guys who make a lot more than a professor. One of our traders tapes the quarters he wins above his his desk to remind people of his victory.

Carl writes:

Saying "$100 isn't much different from $1" is patently absurd. Like you say, $100 might pay for a nice dinner with one's family. $1 might pay for a 1.5 ounce bag of potato chips.

Zac Gochenour writes:

@Carl, I don't mean there's literally no difference. I just mean in relation to the reputation that is already at stake, it conveys very little extra information about your level of certainty.

@David, I agree with Bryan with some remarkable level of consistency. Since most of the bets he offers are at even odds when I think he should be offering much better odds, there hasn't been a bet he's offered that seems like even a remotely good idea to take the other side of.

Tracy W writes:

I just mean in relation to the reputation that is already at stake,

But how much is at stake in most bids? Yes, if you've made a career on saying one thing, then bidding on that is putting a lot on the line. But if you're a pundit that covers a lot of areas, will your reputation be harmed that much by being wrong here and there? A bet that most of your readers won't have heard about, and a chunk of the remainder will vaguely think "oh, yes, Smith, I've heard that name somewhere before".

Meanwhile $100 is $100, a nice meal out. That is something at stake.

Carl writes:

With all due respect, I think Zac and rapscallion are arguing just to argue. Their posts are premised on the assumption that $100 is of little worth to economics professors, which I think 99% of professors would tell you is false.

Comments for this entry have been closed
Return to top