Must you bet your views?

A reader asks:

How about some comments on the refusal by Krugman to bet some of his Nobel money against Mankiw?

Put aside Krugman and Mankiw and let's consider the issue in the abstract.  Bryan Caplan believes that scholars should be ashamed if they do not publicly bet their views.  In contrast I fear this requirement would become a tax upon ideas.  How would you feel about an obligation (if only a moral one) for scholars and commentators to publicly reveal the content of their investment portfolios?  Those portfolios are their real bets.  Yet I still favor the privacy norm and I should note that Bryan never has (nor need he) revealed his portfolio to others at GMU, much less to the broader public.

Let's say that I, as a prolific blogger, express opinions on hundreds of economic policy topics, often involving either explicit or implicit predictions.  Then say that hundreds of people wish to bet with me.  Can I not simply turn them all down as a matter of policy and practicality?

If you're wondering, I practice "buy and hold and diversify," with no surprises in the portfolio and a conservative ratio of equity purchases.  But those investment decisions don't necessarily reflect my views on any given day.  I think it is intellectually legitimate (though perhaps not always prudent) to engage in mental accounting and separate those two spheres of my life.  I change my mind lots of times, on many economic issues, but does that mean I have to become an active trader?  I hope not and I'm not going to.

On long-run economic growth I'm still an optimist, though I am increasingly uncertain as to how much extant firms will capture those gains.  On the short run issue at hand, I am fully with Mankiw, and Megan McArdle, in very much doubting the "rosy scenario" emanating from the Obama budget process. 

Addendum: Robin responds, Bryan responds.


Did Mankiw issue rosy scenarios when he was in the Bush administration? Or if Bush did, did Greg break ranks to warn everyone?

You know, some in the right wing blogosphere have been down on Obama for being pessimistic, for 'talking down' the economy. When Obama issues which is for me a fairly conventional Presidential estimate (they are ALL rosy), they turn smartly on their heels to fault him for that. It's crazy. It seems Obama can neither tell us that the economy is in trouble, or that it is recovering.

(Finally, a good "Fooled by Randomness" reader would not get excited by any GDP estimate, because he knows they are all just guesses.)

Also, it seems to me that the blog post here is taking an idea to an extreme that nobody is proposing. Of course you cannot take the time to bet on everything you think. If you spent all your time thinking about your certainty on everything (and thus the odds you'd accept), you'd not think of anything useful any more.

However, in occasional cases of disagreement where both parties spend a lot of time, I don't think it's a problem to bet in issues where there's pride at stake. Obviously in my world, if you do not start to negotiate for a wager in those situations, you must not really believe what you say.

I'm with the other commenters who think that at least a gentleman's bet is actually required. Maybe not of every scholar in every field, but public intellectuals who vociferously attempt to influence public policy? Damn straight they ought to at least put their pride where their mouth is.

A bet is a signal of the strength of your belief. At some times that is a worthwhile signal to send.

Even considering this idea in the abstract, you have to remember that this isn't a simple disagreement. Krugman titled his post "Roots of evil" and said that Mankiw's views displayed "more than a bit of deliberate obtuseness." If one person, with a decent amount of disposable income, has views that strong, what possible objection could he have to defending them with a small wager? Had Krugman merely expressed his opoinion that Mankiw was wrong, I doubt the wager would have been proposed.

Yeah. The abstraction here does not seem that useful. No, an academic is not always required to bet. But sometimes he probably is.

To hazard an analogy: Is a man required to "defend his wife's honor"? If that is ALWAYS the case, he would be required to deck the supermarket clerk who says the woman is wrong about the price of potatoes. Which is obviously wrong.

But I think it's true that there are SOME cases in which a man can reasonably be expected to intervene.

To risk another analogy: I think that I can opine about sports without being expected to back said opinions with a bet. But if I walk into a bar, see a guy in a Dallas Cowboys jersey watching a game, and say, "God, your team sucks. Your quarterback sucks. Your receivers suck. They are so obviously going to lose, you must be freaking obtuse to think otherwise. Plus your coach is a liar. Your players are all liars. So much so that I believe that they are probably trying to ruin the entire sport of professional football. It has become so bad and so obvious that they are liars, even their fans are liars. Or stupid."

So maybe the guy with the jersey responds, "Wow. You seem pretty sure of yourself. Care to make a wager?"

The obnoxious guy, I think, would be obliged to accept that wager. Maybe not bet his family's house. But a wager? Yes. Even a dollar. Or a beer. He absolutely must accept.

And for all his brilliance, Krugman is kind of obnoxious. Which is good. Great even. But it does obligate him to take the wager.

Also worth noting, the last words of DeLong's post:

"And that is certainly the way to bet."

Well, if it certainly is the way to bet...

Hasn't Tetlock already proven that experts are bad at predicting things in their field of expertise -- worse on average than reasonably informed laypeople? So why make them lose their money to prove what we already know?

They are not talking about a broad thing like "the stimulus." The discussion is about Obama's assumptions as to the growth rate. His budget assumes it will be X percent. Mankiw says, "No it won't." SO you wait until the year(s) in question and see if the administration's assumptions held.

Hey, Billare, care to bet that Krugman tarred Kling as a racist? Here's the column in question.

Just to throw in a mild monkey wrench here, but it happens to be the case that some branches
of various some great world religions consider gambling to be a sin. Hence, there is at least
the possibiliity of being at least vaguely morally repelled by the whole practice.

I am one of those loud-mouthed academics who shoots off his mouth in blogs, although I generally
avoid making very specific forecasts of the sort that are particularly bettable, such as "The
Dow will be above 10,000 by the end of 2010!" I am all too aware of Keynesian uncertainty and
I have a sort of irrational feeling that if I bet on something, that very act will somehow make
it not come about, even if it looked reasonable at the time I made the bet.

But, more generally I just find myself getting really turned off when somebody pulls this old
"you wanna bet on it?" line. It seems like a) a kind of half-baked macho exercise, and b) a way
of actually avoiding discussing the issue in question ("to heck with all this talk, put your money
on the table!"), not to mention that I have better things to do with my money.

I recognize that there is a certain pro-betting attiude in the famous Mason lunch crowd. It is
not just Bryan Caplan, whose views are well known on the matter, but I think Robin Hanson probably
adds to it, given his faith in prediction markets. I grant these often do well, and have even published
some papers showing this (in the journal I edit, JEBO, not by me, but by others, including Robin).
However, I also know there are some cases where they do horribly, with the ones for who is to get
the Nobel Prize in economics each year being a prime example.

One pair of academics who did bet publicly were Julian Simon and Paul Ehrlich, who bet in 1980 over
whether the prices of a set of commodities would be higher or lower in 1990. It was public and non-
trivial, $10,000. Simon's bet on lower prices won, and Ehrlich paid up. I have no problem with others
engaging in betting if they want to (although I think people who go to Las Vegas or anywhere else to
bet in setups with negative expected returns are just fools, and many of them truly pathetic ones), but
I just simply do not like and never have done it even in the zero sum game setups (although I have played
poker sometimes, a different matter somehow, a game of skill after all), not on sports, not politics, not
on economics, although I will not claim the morality argument itself. More just some kind of gut reaction.
Don't wanna and you can't make me.

"Let's say that I, as a prolific blogger, express opinions on hundreds of economic policy topics, often involving either explicit or implicit predictions. Then say that hundreds of people wish to bet with me. Can I not simply turn them all down?"

It wasn't quite like that. It went like this: Krugman accused Makiw that he knows that Krugman is right, but he refuses to say so because he is a Bushie, and of course Bushies lie, steal, and are pure evil. If you are Mankiw, waiting to see if the facts refute Krugman is a lose-lose situation. If they do not, Krugman gloats that he was right and you of course knew that. If they do show Krugman is wrong, he could still say that at the time everyone who was not an evil Bushie believed that he was right. The only way for Mankiw to show that he did not believe Krugman is right at time t zero is to propose a bet, because, presumably not even evil, stupid Bushies would bet that the water is not wet. Once Mankiw does that, what Krugman does is irrelevant: if he takes the bet, he proves that he truly believes what he writes, as stupid and absurd that may be. He did not take the bet, showing that not even Krugman takes what Krugman writes seriously.

Phil P,

Mankiw made the bet parameters quite specific:

Team Obama says that real GDP in 2013 will be 15.6 percent above real GDP in 2008. (That number comes from compounding their predicted growth rates for these five years.) So, Paul, are you willing to wager that the economy will meet or exceed this benchmark?

Considering the language that Krugman used, he would take a gentlemen's wager if he had any backbone whatsoever. Hume, above, explained the refusal- Krugman refuses to put himself in a position where he might have to admit he was wrong.

Tyler, basically what this comes down to is that it's time to pick a side. When Delong, Krugman et al. make baseless and personal attacks on reasonable, honest men like Mankiw, and then refuse to wager on the matter in dispute, do you call out the former for their crassness and intellectual dishonesty? Or do you join the chorus?

I have a feeling you'll be forced to choose sooner rather than later.


Don't we tend to think that costly signaling in a prediction market will tend to yield better information than cheap talk, with no ability to measure/weight probabilities of certainty?

I'd somewhat agreed with the implications of some of the above posts. This feels you're prioritizing the interest of the guild of scholars over a scholar's interest in truth.

Krugman is a cad. But you knew that already.

If I know I'm right with a great deal of certainty, I no longer consider it a bet -- it is a money-making opportunity. I think Mankiw should offer this bet, erm, money-making opportunity as a futures-style contract rather than an over/under.

oops, dropped an italics tag

To me, the value of knowledge is in predictive value. If it doesn't help you predict something moving forward then all your knowledge isn't helping you do anything more than guess. So absolutely I think that someone's opinion simply carries more weight if they are actually willing to make an actual prediction- honestly to me the funds involved are pretty much trivial. Just being willing to go on the record in a way that can be looked back on and examined is what matters. Betting just ensures that someone else will go back and check on the results.

This discussion assumes that public figures believe their own bullsh*t. Which, I doubt. Do you really think Rush believes all of the nonsense he spouts?

Bernard Yomtov is quite correct about the disjunction between personal risk aversion and the quality of ideas. And it is not necessarily just a matter of risk aversion. Imagine an academic who is a recovering gambling addict or has a moral aversion to gambling.
The ideas of such people, on Bryan Caplan's view, would be discounted ab initio by their refusal to take a bet. A typical economist's conceit to ignore the forces that shape preferences.

That Krugman does not accept Mankiw's offer of a bet does not mean he does not stand behind his own ideas, nor does it mean his ideas should be discounted. It only means he is not interested in a bet with Mankiw about them. Any other conclusions are nothing more than speculation.

Maybe the market for ideas would work better if people could sell their ideas if they change their mind. For example you make a $20 bet that there will be positive economic growth in 2010. A month later you doubt that, so you sell your bet to someone else for $15 dollars.


I am not against prediction markets. Just noting that they do not always do all that well. Not clear that the
ones for Nobels really serve all that much of an important purpose other than fun and games. Don't think we
need an alternative institution for that one. The Nobel committees will do their work, and that will be that.

I probably should not say it, but it looks like a comment of mine got censored. Maybe just as well as I used
bad language and was not very nice to some folks. So let me rephrase my point more carefully.

I think that the specific case of Mankiw versus Krugman has some peculiarities, which in the end exhibit the
limits of this betting business as particularly useful for intellectual discourse. So, Mankiw rather damaged
his credibility back when he was CEA Chair under Bush by making statements that many viewed as contradicting
things one could easily find in his textbooks. Krugman used that to go over the top, as he often does, in
engaging in excessively hostile and personalistically nasty remarks about Mankiw in the context of their debate
over how one should or should not do a stimulus plan. Mankiw responded with his bet challenge, to which Krugman
has not responded.

While I think that Mankiw had good reason to be annoyed with the degree of personal unpleasantness issued by
Krugman, I think that his bet was far from as straightforward as it appears, and that for this reason it is
completely justified for Krugman not to respond to it at all, even if one thinks that issuing challenges to
to bet to people with whom one is having an intellectual debate with is just hunky-dory. First of all, Mankiw's
bet is based on the assumption that somehow Krugman is arguing that the specific Obama stimpack is going to "work,"
and that the test of this is if the economy peforms at least as well as the projections made by the Obama team.
There are several problems with this.

1) Krugman has strongly argued that the Obama stimpack is way insufficient. This is of course philosophically
the complete opposite of Mankiw's position, who says one should not be doing a spending-driven fiscal stimulus
at all (or not much of one). However, the implication of Krugman's position is that he does not expect the economy
to do as well as Obama is predicting. Why would anybody expect him to take up such a bet, given this? It is simply
ridiculous on the face of it.

2) Even if one does not think that the stimpack is insufficient, there is plenty of reason to expect the Obama
projections to be overly optimistic. One is simply the general tendency of presidents to issue "rosy scenarios."
The history of that is just way too long. And the other is that many have been complaining that Obama and his
team have been actually damaging the economy by too much grim talk that is inducing fear and thus damaging the
economy. Rather, and I think his team was playing this during the past week, Obama should be being optimistic to
inspire hope and optimism so as to help bring about a self-fulfilling prophecy of a better outcome along the lines
of since we "have nothing to fear but fear itself," let us not be so fearful and engage in happythink.

3) Not unrelated to the previous point, which suggests that an outcome somewhat less than the Obama projections
is more likely, given their politically driven overoptimism, there is the matter that what constitutes "success"
of the stimpack is highly arbitrary. The basic argument has been do spending-driven fiscal stimuli work or not
work? So, here we are going to have a fiscal stimulus, one that Krugman thinks is too small, but which Mankiw
thinks is too big. How do we know if it is "working" or not? Mankiw implies that it must meet the overoptimistic
Obama team projections. But, I would say that if the outcome is not much below that, the case would be very strong
that it had "worked," or at least not done too badly. Somehow demanding that Krugman pay up if the outcome is 0.1%
less than the Obama projection just seems pretty ridiculous.

Bottom line is that this was a phoney bet proposal that would establish nothing one way or the other, and was stacked
in Mankiw's favor. Such bets really need to be much more clearcut and straightforward, a la the Simon-Ehrlich one of
the 1980s on commodity prices. Krugman is no fully justified grounds in not even giving it the time of day.

A more general comment, which extends my earlier ones, is that I think betting is fine in voluntary, zero-sum games
of skill, such as poker (although of course there are plenty of suckers who get taken for their money in such games).
But as a method of intellectual discourse or debate, challenging someone to a bet is basically dirty pool. It is
a macho stance that vaguely harks back to the dueling we had in the past. It cuts short discussion. "Bet or shut up!"
The comments made by many here have simply reinforced my view of this, with people talking about how Krugman has no
"backbone," is clearly not a real man, blah blah. This is just a sorry spectacle, although I would also grant that
some sub-groups, such as some of the members of the Mason lunch crowd, might view this as a form of poker, wherein
people are regularly making bets with each other in a friendly manner about various things, a sort of localized
version of Robin Hanson's prediction markets (maybe just between him and Bryan Caplan, given Tyler's apparent lack
of interest, although maybe they can get Tyrone into it sometime, :-)).

As it is, my own experience has been that when someone challenges me to a bet about some matter of economics or
politics, it is because they have run out of arguments, and they are trying to gain some macho edge by their move,
which, as I have already said, I never accommodate or respond to favorably, spineless wimp that I am...

This whole idea is dumb. Firstly (and trivially) some people are just not comfortable with gambling. To be condemned as having no backbone for this is unfair.

Secondly, I have sometimes bet against my truly held point of view as a sort of hedge on my opinion. If I am right I lose the money but have the inner glow that comes from being proved right and if I am proved wrong the bet pays off and so I have some cash to cheer me up.

As far as the specific Mankiw/Krugman challenge is concerned, it is puerile at best.

Mike Moffat,

I find your argument that any prediction not accompanied by a bet is
"noise pollution" to be seriously unconvincing.


The problem is that the cutoff Mankiw proposed was the overoptimistic
projection made by the Obamoids. I would suggest that most people
would say "the economy recovered" (if not fully sufficiently) rather
than "it went south" if it showed half the growth rate of the cutoff
Mankiw proposed. This was a severely stacked bet.

Also, it has been brought to my attention that the argument between
them had more to do with unit roots precisely than the Obama fiscal
stimulus, although the latter was clearly behind the precise debate.
On this matter of unit roots, frankly, and despite his personal
unpleasantness in the matter, Krugman is simply correct. The current
literature in time-series macroeconometrics no longer supports the
view that Mankiw pushed. This conclusion was not driven by people
who foresaw that Mankiw would want to use this now-out-of-date view
of unit roots to criticize the efficacy of a proposal by Obama. It
has been around for quite some time now.

Well, if the only problem is cutoff growth rate, Krugman can make counter-offer which he thinks attainable (say, 8%). But even then, the risk structure remains unfavorable to Krugman, as I described in my previous comment. So unless Krugman is risk-lover or the Schadenfreude to beat Mankiw is so great, it will be irrational for him to accept the bet, whatever the cutoff rate would be.

bzzw thank you! cialis <a href=>cialis</a> mf612

Comments for this post are closed