Scott Sumner  

Why Bryan Caplan almost always wins his bets

PRINT
My Response to Scott Sumner... Why I Think I Win My Bets...

[This is aimed at clarifying my previous post, which David commented on.]

Bryan almost always takes the consensus view in any bet. The view a betting market would take, if one existed. When I first hear about a new Caplan bet, I almost always agree with him, and think that he's likely to win the bet. The other side of the bet puts too much weight on their own subjective view. They forget that they are nothing special, and that the smart bet is the consensus view, not the particular hypothesis that they have fallen in love with, because it's theirs.

Most people (perhaps including me) tend to overweight the probability that their views on religion, politics, economics, sports, and art are correct. (On art, by "correct" I mean that the work of art is likely to be perceived that way in the distant future.)

In terms of my area of specialization, monetary policy, I take the following view:

I do not believe that the Fed should set interest rates where Janet Yellen thinks is appropriate, or where John Taylor thinks is appropriate, or where I think is appropriate. Rather interest rates should be set where the market thinks they are most likely to lead to good macroeconomic outcomes. Yellen and Taylor (and me?) do have an important role to play in figuring out what "good macroeconomic outcomes" means.

There is one possible flaw in always marking our beliefs to the market consensus. It's been pointed out that if everyone believed the EMH made it impossible to consistently beat the market, then no one would do the research necessary to make the EMH true. Thus we need lots of individual actors trying to figure out where stocks should be valued, and forming opinions independent of the current market price. Otherwise the EMH does not work. But they should also be wise enough to know that they are nothing special, and hence they should advise their mother to invest in index funds, not their own hedge fund. That's what I mean by realizing that you are nothing special. You are merely a worker ant making the EMH more true.

In macro, it's important for people like me to always search for the truth, and reach conclusions about economic models in a way that is independent of the consensus model. In that way, I play my "worker ant" role of nudging the profession towards a greater truth. But at the same time we need to recognize that there is nothing special about our view. If we are made dictator, we should implement the consensus view of optimal policy, not our own. People have trouble with this, as it implies two levels of belief about what is true. The view from inside our mind, and the view from 20,000 miles out in space, where I see there is no objective reason to favor my view over Krugman's.

In the Middle Ages, people believed their own religion was the one true faith, and used public policy to try to make most of society adhere to this one faith. They could not understand the concept of the view from 20,000 miles out. In the enlightenment (after Copernicus), philosophers realized that we have no privileged position to decide the one true faith, and that public policy should not favor any one religion.

I also believe that public policy should not try to reshape our population more along Republican or Democratic lines, as we don't know which party's views are correct. Thus immigration policy should be neutral, except about characteristics for which there is a broad consensus that they are desirable or undesirable. (Say Nazi views, or totalitarian communist views, which are clearly wrong.)

PS. There was some confusion about my recent denial of personal identity. Most people believe there is a "you" that possesses a brain, and a "you" that possesses a body. I don't, I believe there's merely a brain and a body, no "you". This confusion shows up in discussions of free will. In this recent Atlantic article, the writer discusses the implication of there being no free will for optimal criminal justice policy. I agree with the writer about there being no free will, but he fails to understand that this view has no implications for criminal justice policy. The optimal policy is the one that maximizes aggregate utility. But if you look at any economic model of crime you'll immediately see that the model works exactly the same whether free will exists, or it does not exist. There is no special "FW" variable in the model's equations, which is one if free will exists, and zero if it does not. Since the free will assumption has no practical implications for behavior, it's best to assume that free will does not exist, which also means that there is no personal identity, just a brain spinning out one thought after another, according to the laws of science (plus chance?). I know others won't agree with me, I merely provide this to explain my view.


Comments and Sharing






COMMENTS (31 to date)
mico writes:

"In macro, it's important for people like me to always search for the truth, and reach conclusions about economic models in a way that is independent of the consensus model. In that way, I play my "worker ant" role of nudging the profession towards a greater truth."

In a democracy, it's important for voters like me to always search for the truth, and reach conclusions about political questions in a way that is independent of the consensus model. In that way, I play my "worker ant" role of nudging the country towards a greater truth.

Garrett M writes:

I think that changes in consensus views (trends) are important to know as well. The wisdom of crowds is a powerful force, but to the extent that new information is assimilated with any degree of lag (doesn't have to be much), the directionality of prior updates is useful information.

anomdebus writes:
But if you look at any economic model of crime you'll immediately see that the model works exactly the same whether free will exists, or it does not exist. There is no special "FW" variable in the model's equations, which is one if free will exists, and zero if it does not. Since the free will assumption has no practical implications for behavior, it's best to assume that free will does not exist

I think this is missing a "in economic models of crime" at the end of that quote.

Philo writes:

"I do not believe that the Fed should set interest rates . . . where I think is appropriate. Rather interest rates should be set where the market thinks they are most likely to lead to good macroeconomic outcomes." But that is where you think is appropriate! You are using a "prediction-market rule" rather than a Taylor rule or whatever. (And you are right to do so!)

By the way, there never will be a prediction market in whether God exists, or in whether (some sort of) utilitarianism is true, or in any other philosophical issue; in philosophy you can consult the arguments of other people, but ultimately you have to rely on your own judgment.

Oliver writes:

We're still in a bit of a philosophical black hole here. You said:

I do not believe that the Fed should set interest rates where Janet Yellen thinks is appropriate, or where John Taylor thinks is appropriate, or where I think is appropriate.

That simplifies to: I do not think it is appropriate to set interest rates to where I think it is appropriate to set interest rates. Well, that can't be right.

From your comments on the previous post and David's response, it seems that the intended qualifier to that sentence is: "just because I say so."

But if that's the point, surely you can get there without feeling the need to argue that "objective" and "subjective" are the same thing. And it would perhaps be helpful if you could quote somebody---anybody!---made more of flesh than straw who makes this kind of argument. I can't recall ever hearing one.

Scott writes:

People have trouble with this, as it implies two levels of belief about what is true. The view from inside our mind, and the view from 20,000 miles out in space, where I see there is no objective reason to favor my view over Krugman's.
There are at least two levels of "truth". Comparing 20,000-miles-out views, the Resource-Patterns Model of Life suggests one view may prove viable, the other fatal. See my quiz (required of all my zero students) and answers.

We human beings (living things on level N in a process of life in levels) organize ourselves through our efforts and discussions to find the objective facts which may enable greater flourishing at a higher level (level N+1).

bill writes:

I like Sam Harris's book on this topic.

RPLong writes:

This is a good post because it is highly descriptive of some of Sumner's most controversial beliefs. Getting a look at the debatable assumptions behind his ideas is a boon to understanding the whole picture.

As an aside, internet commentary (i.e. in the comments) seldom covers the so-called "free will theorem," which I consider very important to the discussion. I offer a link to those who might not have seen it before.

Scott Sumner writes:

Mico, Exactly.

Garrett, That makes sense.

Anomdebus, I don't think the free will hypothesis has any implications in any area of life.

Philo, You said:

"But that is where you think is appropriate!"

Again, you are confusing two levels of belief. Recall my analogy with the stock investor. At one level he things his hedge fund is best, but at a higher level he thinks him mom should buy index funds. At one level I might think a 3.5% fed funds rate is best, but at another level I defer to the market view of where interest rates should be. I am merely one person in that market.

Is it possible that the market might not think the market forecast is best? I suppose anything is possible, but that would be really weird.

Part of the confusion here is that our language is built around the normal person's (faulty) view of reality, not my (Rortian) view.

I agree with your second point, which is related to my claim that the market can't decide what "good macroeconomic outcomes" means, economists must decide.

Oliver, See my response to Philo. Part of the problem here is the English language lacks terms to describe my view of reality. Belief operates on many levels, as I tried to indicate in my post.

Thus at one level I believe monetary offset is true, at another higher level I think it's just one of many macroeconomic hypotheses. Ditto for libertarianism.

jonathan writes:

So if we apply this view to Trump, would you agree that his probability of winning the presidency is on the order of 30%? (per betting markets). And since this probability has increased significantly over the last year without obvious corresponding movements in world markets, can we infer that his election would not be particularly problematic for the world?

I guess we could express this as a bet, e.g. of some particular disaster if Trump is elected president. I would probably bet that the odds of any particular disaster are not significantly different from the base rate (maybe slightly on a few topics).

Josiah writes:

Scott,

Bryan based his bets on a consensus of *experts* not a consensus of everyone. He wrote a whole book about how you shouldn't rely on a consensus of everyone. And in any event, neither an expert consensus or a general consensus is the same as a market consensus.

Also, you say that we shouldn't let immigrants' political views affect our views on immigration because we don't know which policies are correct. Then you say we should based our criminal justice policy on utilitarianism. Leaving aside how we are supposed to do that if we don't know what the consequences of different policies are, have you considered that immigrants may not be utilitarians?

anomdebus writes:
Anomdebus, I don't think the free will hypothesis has any implications in any area of life.

I don't see how you showed that with the quotation in question. You are assuming you know what the economic model is under those two different scenarios, then saying they are the same.

I also do not agree that it necessarily has no impact on criminal justice policy. I believe the use of the insanity defense shows that perception of free will changes criminal justice policy. Mentally incapacitated people are not treated the same as people with similar criminal activity.

Could we really keep all the same policies, while transitioning from a majority believing in free will to a belief in lack of free will? I think that is a big assumption that you would be responsible for, free will or not.

Philo writes:

"At one level he thin[k]s his hedge fund is best, but at a higher level he thinks hi[s] mom should buy index funds." So he will advise his mom to buy an index fund, which shows that what you call his "higher (highest?) level belief" is his belief simpliciter, and what you call his "lower level belief" is not really a belief, but just an inclination to believe that is overridden or outweighed by other considerations.

By the way, if forced to choose between the normal person's view of reality and Rorty's view, the sensible person will choose the former!

dlr writes:

Scott,

Let's say you were named Fed Chair, with the unique ability to use a hidden provision to override the entire FOMC if you felt like it. You wanted to implement an NGDP futures market and an NGDP price level target. The entire FOMC politely disagreed and voted to stick with the existing rate-centric hyper-discretionary regime, and one member even commissioned a nationwide poll or perhaps economist/investor wide poll in which over 90% of respondents sided with the rest of the FOMC. Let's even say that something like 2008 started to happen, and the rest of the FOMC was about to repeat its mistake, with economists lining up to support them in letters to the editor. Do you bow to consensus and abstain from acting alone because you are nothing special?

Gordon writes:

"Thus we need lots of individual actors trying to figure out where stocks should be valued, and forming opinions independent of the current market price. Otherwise the EMH does not work."

Isn't this a good argument to place limitations on large investment firms from using computers to make trades that are based on recent price changes? From what you're saying, the more algorithmic trading takes place, the more the EMH will be undermined.

Scott Sumner writes:

Bill, Thanks for the tip.

RPLong, Thanks.

Jonathan, I think that's probably right, although it's hard to be sure. In addition, recall that a 1% chance of a disaster that wiped out the stock market, say a nuclear war, would not noticeably affect stock prices. But it would still be a big concern.

As an aside, I'm strongly opposed to Trump, but not because I think he'll have a dramatic effect on the economy. I doubt even Sanders would move the market noticeably, at least when his probability of being elected moved from 15% to 30%. It would be swamped by other factors moving the markets. Studies show there is about a 2% difference in stock prices depending on whether the GOP of the Dems win, with prices a bit higher under the GOP. With Sanders it might be 4%.

Josiah, I agree that expert opinion is often better than the views of the average person, but I think the market forecast is a bit better still.

I don't think either immigrants or domestic residents are particularly utilitarian, unless you live in the Nordic countries.

anomdebus, Most believers in free will and non-believers agree there is something called insanity. They may disagree with how much control the insane have over their behavior (i.e. the debate over Szasz's views, etc.), but that's not a debate about free will, it's a debate about insanity.

Philo, Then I'm not sensible.

dlr, You asked:

"Do you bow to consensus and abstain from acting alone because you are nothing special?"

Yes, it would be irresponsible of me to do otherwise. Imagine I am behind that famous Rawlsian curtain of ignorance, and was told that some lone economist blogger would be offered the right to overrule the Fed, but we did not know who. Would I think that a good idea? No. So the fact that I know it's me should not lead to a different answer, as I'm not special.

Gordon, I'm inclined to think that that sort of trading neither helps nor hinders, but I don't know enough to have a firm opinion. You might be right.

Alex writes:

"Most people (perhaps including me) tend to overweight the probability that their views on religion, politics, economics, sports, and art are correct"

But you may also be overweighting the view that most people overweight their views.

Its like saying "Knowledge is impossible"
How do you know?

travis allison writes:

Scott Sumner said:

"Thus immigration policy should be neutral, except about characteristics for which there is a broad consensus that they are desirable or undesirable. "

"I don't think either immigrants or domestic residents are particularly utilitarian, unless you live in the Nordic countries."

So my question is why should anyone act according to the broad consensus simply because it is the consensus?

DZ writes:

Gordon said "Isn't this a good argument to place limitations on large investment firms from using computers to make trades that are based on recent price changes? From what you're saying, the more algorithmic trading takes place, the more the EMH will be undermined."

This is no longer _the_ debate over HFT. The market already made its decision in wanting to eliminate algo sub-penny frontrunning. Exchanges like IEX have created the "speedbump", slowing all trades to a uniform execution time by simply including a 350 nanosecond delay (physically using a longer connection wire).

IEX became very popular and applied to become a protected exchange (like the NYSE), but is being prevented by large HFTers (Citadel) lobbying the SEC, claiming the speedbump violates Reg NMS, which requires immediate execution of trades. So now the debate is on the other end. Should government prevent the market from allowing exchanges to take measures to prevent tech front running. The answer to me seems clear.

Philo writes:

In most things, you're admirably sensible (insightful, etc.). In philosophy . . . well, better stick to your day job!

Scott Sumner writes:

Alex, Most people claim they are better than average drivers. How likely is it that this is true?

Most people are religious, and believe their particular religion is the right one. How likely is it that this is true?

I could find similar examples in all sorts of areas of life.

Travis, It depends what you mean by "act". Our society is set up where people make independent decisions such as voting, and then public policies reflect the consensus of the voters. I do not want people to vote according to the consensus view. I want them to think independently.

But if I lose the battle to convince others of my view of monetary policy, it would be arrogant of me to overrule the views of the majority of economists if I were made dictator of the Fed. There is no OBJECTIVE reason for me to believe my view is superior to the consensus view.

People always play two roles; they come up with independent opinions about the world, and they also observe the consensus view, and understand that the wisdom of the crowd implies this view is the optimal forecast.

Thus "belief" is always a difficult concept, as there are two levels of belief in any wise man or woman.

Scott Sumner writes:

Philo, My monetary views are based on the wisdom of the crowds, the EMH. So if I'm wrong about philosophy then I'm also wrong about monetary policy.

DZ, Very interesting, do you have a link?

anomdebus writes:

Scott,
I mention these points because I think it is a weakness in your argument. You say that it has no effect on the outcome, yet insist on assuming it.

I think you would be much better off skipping the whole "is there a you?" for the purposes of eliminating a privileged viewpoint and just stick with "I can't say what your preferences are, nor you mine, therefore neither of us can dictate an optimal policy for the other".

Mike Sax writes:

I came across the idea of no free will in this interesting book by Frank Harvey regarding the idea of counterfactual analysis.

https://www.amazon.com/Explaining-Iraq-War-Frank-Harvey-ebook/dp/B006NYCK4G/ref=sr_1_1?s=books&ie=UTF8&qid=1464373149&sr=1-1&keywords=frank+harvey

Harvey argued that the idea of many Democrats like myself that had Gore been President there would have been no Iraq War is mistaken for two reasons.

1. 9/11 would have forced Gore to basically the same decision.

2. Long term structural factors in American power. For instance , the official policy of the US under Clinton-Gore was also regime change in Iraq.

I don't know if I agree but it's an interesting theory anyway.

Towards the end of the book Harvey revealed that he believes there's' no Free Will.

Turns out there is a whole school of thought to that effect.

http://www.informationphilosopher.com/freedom/problem/

The problem with this theory for some will no doubt be how you assign moral responsibility if you believe that people had no 'free will' for their decisions in the first place.

DZ writes:

Scott,

The full back story is covered by Michael Lewis in Flash Boys.

Here's an article on how the founders of IEX discovered the demand for this service:
http://www.cfapubs.org/doi/pdf/10.2469/cp.v32.n4.6

Here's a more technical article with evidence depicting how HFT is negatively impacting markets.
http://www.nanex.net/aqck2/4661.html

You have to understand, the existence of a market-derived fix to the HFT issue was obviously not expected. Most liberal economists at Cato Institute, CFAI...etc. spoke out against preventative HFT regulation claiming that HFT improved market liquidity blah blah. Now, the evidence is clear that both sophisticated and everyday investors are being fleeced by HFT firms. All these individuals whom spoke out supporting HFT have gone silent on the issue now that it is government regulation itself keeping HFT alive.

The best source of updates I know of is Eric Scott Hunsader at Nanex (market structure application firm). His twitter is @nanexllc. He tweets about the issue multiple times each day.

Philo writes:

I think Bryan Caplan explains that *the market* in EMH is not *the crowd*. And, by the way, Richard Rorty is also not *the crowd*.

Sebastian H writes:

How do you know that the people who think they are special are wrong?

Epistemic humility is good for analysis, but like many insights can be taken too far.

Scott Sumner writes:

anomdebus, You said:

"I mention these points because I think it is a weakness in your argument. You say that it has no effect on the outcome, yet insist on assuming it."

Not at all, I do not insist on assuming it. My argument applies equally well if there is free will, or if there is not. In fact I believe all arguments apply equally well in a world with and without free will. If that were not so, then it would be possible to test for free will.

DZ, You said:

"All these individuals whom spoke out supporting HFT have gone silent on the issue now that it is government regulation itself keeping HFT alive."

I find this difficult to believe. If true I'll post on it, so please send me any evidence you have. Do you have any Cato links where they support regulations blocking this new type of market? Most of the arguments against Flash Boys that I read from free market types were exactly the opposite--the best way to address this problem, if it is a problem, was competition from markets that did not allow HST. That was my position. Did these free market advocates suddenly abandon that position? I have not, so I'm at least one exception to your generalization.

My second question is whether regulation has shut down the market that bans HFT. I have not heard about it being shut down. Has it been shut down?

And if what you say is correct, then Cato was correct to oppose regulation, it's not needed.

Philo, Markets are the optimal way of deriving the wisdom of crowds. Where you don't have markets, you look at the consensus view of the most qualified experts.

I'm not persuaded by Caplan's posts, as I believe the consensus would use the same type of forecasting procedure that he used. Again, when I saw his proposed bets I almost always had the same view. So in that sense there's really nothing very impressive about his success, it's the other people who made foolish bets, like the claim that QE would lead to high inflation. Or they gave foolish odds (as I seem to recall in the unemployment bet.)

Sebastian, You said:

"How do you know that the people who think they are special are wrong?"

Not only do I not know they are wrong, I believe some of them are right. Copernicus was right. Einstein was right. My point was that you are not special by virtue of being you.

Who knows, there's even a slim chance that I'll be right. :)

anomdebus writes:

Scott,
If you believe that, then you erred in bringing it up as a means of support for the idea of non-existence of "you".

Scott Sumner writes:

Anomdebus, The "non-existence of you" argument is also unessential to my post on the two levels of belief.

I added it here merely as a postscript.

DZ writes:

Scott we are on the same page but I didn't communicate my last post well. I, as well as the classically liberal economists I cite, have always been against regulation in this market for the reasons you stated. However, several have also supported HFT as improving markets. From what I've read, I've observed that HFT does the opposite, hence why there's strong demand for markets to weed it out. It seems unlikely that economists who supported HFT as helpful (even if they denounced regulation) would turn around and support action which eliminates HFT (even if that action is _removal_ of existing regulation). Either way, the spirit of the NBBO regulation (Reg NMS) for protected exchanges is not to stop markets from weeding out HFT. The SEC lobbying supporting that regulation is telling.

Comments for this entry have been closed
Return to top