Scott Sumner  

Economics must be harder than it looks

PRINT
What Does Public Schooling Tea... Value of Self-Rated Health Ble...

Mark Sadowski sent me this paper from Paul Davidson, which he thought was hilarious:

There are two different major economic theories that attempt to explain the operation of the money using, entrepreneurial economy that we call capitalism and its financial markets. The first and most widely recognized one, and the one that I believe most members of this audience accept as correct is the classical economic theory which is sometimes referred to as the theory of efficient markets, or "mainstream economic theory". Efficient market theorists claim that if economics is to be a science it requires rigor, consistency, and mathematics. Nobel Laureate Paul Samuelson has explicitly added one additional characteristic-- an axiom that most efficient market theorists implicitly assume- namely, economists must accept the ergodic axiom in their models if economics is to be a rocket science on par with physics, astronomy, and chemistry. I will explain in a moment what this ergodic axiom ( a term Samuelson borrowed from statistical mechanics) presumes regarding knowledge about the future. But first I raise the question that if efficient market theory possesses the characteristics attributed to it by its advocates, then how is it possible that efficient market theorists did not foresee the financial crisis that started in 2007-8? Moreover how many risk managers in the audience who had to put their employers' money where there computer model mouths were, saw the financial collapse of 2007-2008 coming? The mantra of this efficient market theory is that enlightened decision makers in free financial markets will always know the future outcome for any decision made today. These decision makers, therefore, always pick those choices that provide the highest possible future returns. Consequently, free markets are efficient in the sense of allocating capital to its best (most profitable) use. Any government interference in these efficient markets will, therefore, always produce a less than optimal outcome. In other words, interventionist government economic policy is the problem, while the free market is the solution.


Whether they declare themselves Monetarists, Rational Expectation theorists, Neoclassical Synthesis [Old] Keynesians or New Keynesians, the backbone of all mainstream theories is the efficient market analysis model where it is presumed that correct information about the future exists today and this information can be obtained by decision makers.


Surely he can't be serious? This reminded me of those Paul Krugman posts where he accuses people like Fama, Barro and Cochrane of making inane comments about the multiplier and/or AD. On the other hand I did a post showing that at least some of these comments were probably much less inane than they appeared to be. Multiple interpretations are possible. Are there alternative interpretations for the Davidson comment? I toyed with the idea that he wasn't implying they ought to be able to predict the timing of the crisis (which the EMH says is impossible) but just the fact that the system was susceptible to such crises, and needs to be fixed. Now even that interpretation would have been wrong (Neil Wallace did warn us) but slightly more defensible. But each time I reread the quotation I see a much bolder and more indefensible argument being made. What do you think?

I'm also reminded of a famous multiple choice question on opportunity cost that was asked of 200 economists at an economics conference. I thought the answer was obvious, and so did Alex Tabarrok, but the actual responses were equally distributed among the 4 choices. Then Tyler Cowen did a post arguing another answer was plausible. Then someone wrote a paper arguing that all four possible answers were defensible.

I'm still not convinced, but which of the following two options seems more plausible?

1. About 80% of the profession, including Tyler Cowen, doesn't understand EC101.
2. Economics is trickier and more susceptible to multiple interpretations than it appears at first glance to me, and to Paul Krugman.

PS. In case Noah Smith is reading this post; that was a rhetorical question. I am not actually claiming Tyler Cowen doesn't understand EC101. He literally wrote the (text) book. Last time I posed this sort of rhetorical question Noah insisted I WAS calling Izabella Kaminska a phony. Lots of other very smart people agreed. On the other hand lots of other equally smart people thought my actual meaning was obvious. Which sort of proves my point, doesn't it.

PPS. I'm also puzzled by Davidson's claim that rational expectations is equivalent to perfect foresight. Or did I misunderstand that passage as well?


Comments and Sharing





COMMENTS (23 to date)

'This is where I came in';

A path-dependent sequence of economic
changes is one of which important influences
upon the eventual outcome can be exerted by
temporally remote events, including happen-
ings dominated by chance elements rather
than systematic forces. Stochastic processes
like that do not converge automatically to a
fixed-point distribution of outcomes, and are
called non-ergodic.

That's from Paul David's infamous paper 'Clio and the Ecohnomics of QWERTY'.
Economics HAS TO BE harder than one thinks, to get a theory of market failure as obviously false as QWERTY accepted.

Yes, non-ergodicity, I remember it well.

Enial Cattesi writes:
I'm still not convinced, but which of the following two options seems more plausible?

Is both an accepted answer? If not, then 1. I mean there are economists who think that economic crises can be caused by too much savings! How can there be too much savings?!

brendan writes:

Note that this guy bashes Samuelson for EMH-belief, and thinks the EMH, if true, ought to have predicted the crash; and Samuelson, in turn, charged Friedman with the same (similar?) error: "Today we see how utterly mistaken was the Milton Friedman notion that a market system can regulate itself,"

But in 2004, Friedman himself admitted that he didn't believe even a weak form (much weaker than described by Davidson) was precisely true:

"You don’t have to believe it. I don’t believe it. We all know the market is not efficient in a descriptive sense. But that doesn’t mean that the efficient market is not the best approximation if you don’t have anything else to use. …Warren Buffett proves that there’s not an efficient market, and yet Warren Buffett is what makes the market efficient, and both statements are right. If the market were 100% efficient, nobody could make any money making it efficient, and then it wouldn’t be efficient again. So in a way it’s self-contradictory to suppose that there really is an efficient market."

Since Friedman writes to be understood (whereas davidson writes to intimidate and impress), I think Davidson's misunderstanding of his opponents beliefs is by design.

MikeDC writes:

I used to start teaching opportunity cost with the multiple choice example (the right answer is $10!).

Anyway, what it tought me, and what's evident here, is that economics is (or isn't) a science only to the extent that the scientists can agree to use the same tools.

Here, the tools are conventions of language. If one physicist used inches and another used meters, their conversations about distance wouldn't make any sense.

Economists need to be more precise in their language. Pick a convention and stick to it. Deviations from the accepted meaning should be held up to the same scientific scrutiny any other test is held up to, and should be utterly dismissed when fallacious.

In terms of good teaching, the opportunity cost example is a great one because you can go through and highlight how choosing slightly different meanings of opportunity costs leads to different "correct" answers, and then you can go through and eliminate the meanings that are more inconsistent (and actually employ other concepts).

To the extent we can do that, create those conventions, and build on them, we build a science.

To the extent we play with those meanings and give credence (or extend the umbrella of membership in our discipline) to an obvious charlatan like Davidson who doesn't take seriously the importance of those meanings and conventions, the whole enterprise suffers and becomes that much less a science.

Thomas Boyle writes:

The term "efficient market" was an unfortunate word choice, because "efficient" has too many possible meanings.

The closest I can come to explaining "efficient market" to laypeople is to say, "no matter how much information you collect, you cannot reliably predict the direction the market will move next". This isn't technically correct (markets move up slightly more often than down), but it certainly addresses Davidson's core misunderstanding, which is to the effect that efficient market practitioners should have foresight. In fact, an efficient market is one in which foresight is impossible!

Politics Debunked writes:

re: "I'm also reminded of a famous multiple choice question on opportunity cost that was asked of 200 economists at an economics conference."

I'd suggest thinking about an altered version of that example. If the cost of the Dylan tickets were $50 (with a "benefit" of $50 ascribed to seeing the show), according to the "right" approach to the problem then the net benefit of seeing Dylan would be $0. Presumably then when offered a free Clapton ticket, if you value seeing Clapton at only $1 then in theory it is a net benefit to see Clapton instead of Dylan. The question is whether people would actually think that way. Since you value Dylan 50 times more than Clapton, would you really choose Clapton?

The issue comes down partly to net vs. gross opportunity cost, and when making a decision like that I'd suggest that many people would focus on the gross opportunity cost to evaluate the choice (while others would use net). I suspect that if people are confused about which to use then if they aren't spending much time on the question they might err by somehow mixing the two concepts together and winding up somehow with one of the other choices.

Mike Sandifer writes:

Scott,

I read and listen to many economists and took a few econ courses in college years ago. I also discuss economics with those who've also taken classes and even some with degrees in the subject on occasion. For what it's worth, my impression is that few, if any of them understand even the fundamentals of economics well.

As I've pointed out before, not one of the many I've talked to who've merely completed some undergraduate work even have a clear idea of what an economy is. In fact, they don't even venture a definition. I haven't asked such questions of degree holders, but maybe I should in the future.

I think it says a lot when people like Krugman will sometimes argue for protectionism to counter Chinese beggar-thy-neighbor policies; when Stiqlitz argues that fiscal stimulus is even needed when there is no output gap, lest an economy lose velocity; when many, many conservative economists continue to predict that high inflation is around the corner, or worse, that "loose" monetary policy leads to deflation; when there's still an active school of thought that changes in the money supply have no real effects on output; etc., etc.

My impression is the subject isn't taught well, and perhaps that's because so few of the people teaching it really understand the fundamentals themselves.

Radford Neal writes:

The opportunity cost question seems to me to be fundamentally mis-specified. It fails to say what you would do if you didn't go to either concert. Suppose that the answer to that is that you'd do some consulting work, for which you would charge $1000. (And let's assume that your client will pay you for as many hours of consulting as you're willing to do.) Going to the Dylan concert must be worth $1010 to you (after subtracting the $40 it costs), since it's stated that that's what you would do if you don't go to the Clapton concert. So the opportunity cost of going to the Clapton concert instead is $1010. Of course, the figure would be even higher if your client is willing to pay even more per hour, so there is no well-defined answer. (And note that the answer will also change if we ascribe a non-monetary benefit or cost to the act of doing consulting work.)


Michael hansberry writes:

The famous multiple choice question does not provide the cost the ticket winner would be willing to pay to see Clapton. There is not enough information provided to answer the question. That the ticket "had no resale value" does not inform us of the value the ticket winner placed on seeing a Clapton concert.

I am not at all sure the writer of the famous multiple choice question understands opportunity cost.

Michael Hansberry writes:

Adding to my previous comment: suppose "On any given day, you would be willing to pay up to" 10 dollars to see Clapton. Then the opportunity cost of seeing one over the other is zero as both concerts are offered at 10 dollars less than one is willing to pay on any given day. (the bargain is the same)


Without the information regarding what one is willing to pay to see Clapton the question cannot be answered.

David C writes:

I think it's always possible to throw out ifs and buts with such questions, but it seems reasonable to assume in the opportunity cost question that Dylan and Clapton are the only options and that both are only available at that one time. Other possibilities aren't typical. I think the 2nd adapted question from the Potter/Sanders paper pretty clearly indicates net opportunity cost rather than gross opportunity cost, but the original probably depends on how you were taught.

The problem, I think, with Potter/Sanders is that they never do the complete calculation for all of the variables. They use the equations:
Bd($50)-Pd($40) for Ferraro and Tyler
and
Bd($50)+Pc($0) as their alternative
but the complete net opportunity cost equation is actually:
Bd($50)-Bc(not given)+Pc($0)-Pd($40)
or in other words:
$10-Bc

For their alternative example involving the movie, they give the equation:
Wb($20)-Pb($0) as their method I to get $20
but the complete net calculation is actually:
Wb($20)-Bm(not given)+Pm($10)-Pb($0)
or
$30-Bm

Also, why Wb instead of Bb?

Kevin Klein writes:

As an independent observation, at least 99% of people with earned PhDs in economics do not understand game theory. They know far less than undergraduates in evolutionary biology or computer science.

So one explanation for why the profession cannot agree to internally consistent definitions of its own concepts is obvious.

Tracy W writes:
The first and most widely recognized one, and the one that I believe most members of this audience accept as correct is the classical economic theory which is sometimes referred to as the theory of efficient markets, or "mainstream economic theory". Efficient market theorists claim that if economics is to be a science it requires rigor, consistency, and mathematics.

I find this totally confusing.
Firstly, I thought classical economics was economics before the marginal revolution, which happened sometime in the 1860s/1870s.
Efficient market hypothesis came around about 100 years later. Basically I thought no one alive today is a classical economist.

Secondly, what on earth is the relationship between being an efficient market theorist and claiming that economics requires rigour, consistency and mathematics? Has anyone come across an economist who claims that economics should be lazy and inconsistent?

Furthermore, one can think about efficient markets theory without involving mathematics (although testing it does require mathematics), and one can think about numerous other economic theories using mathematics even if you think that the efficient markets hypothesis is crazy.

Unlearningecon writes:

It seems to me Davidson is discussing the formal theoretical version of the EMH, which has all of the properties attributes to it, whereas you endorse the informal version floating around the blogosphere, which essentially collapses into the proposition "markets are pretty good at processing information". You're having two different debates.

Thomas Boyle writes:

Unlearningecon,

The formal theoretical version of the EMH explicitly prohibits the possibility of forecasting a crash (or, in general, the future) from the information available in the present. Formally, it states that the price reflects all the information available: the risk of a crash, given available information, is priced in, as is the potential for further gains, given available information.

Since Davidson expects that foresight should be one of the attributes of an efficient market, he simply doesn't understand the EMH.

Thomas Boyle writes:

Unlearningecon,

The formal theoretical version of the EMH explicitly prohibits the possibility of forecasting a crash (or, in general, the future) from the information available in the present. Formally, it states that the price reflects all the information available: the risk of a crash, given available information, is priced in, as is the potential for further gains, given available information.

Since Davidson expects that foresight should be one of the attributes of an efficient market, he simply doesn't understand the EMH.

J Mann writes:

Unlearningecon, I don't recognize any of Davidson's properties as being part of the formal theoretical version of the EMH. Do you have a text or source I could review.

Davidson alleges that:

- "Efficient market theorists claim that if economics is to be a science it requires rigor, consistency, and mathematics." I don't see that this quality aligns to EMH theorists. Do Keysians and post Keynsians (which Davidson posits as the alternative to EMH) reject rigor, consistency and mathematics? I'm most familiar with DeLong and Krugman, and they seem to apply all three.

- "The mantra of this efficient market theory is that enlightened decision makers in free financial markets will always know the future outcome for any decision made today." I think that's completely false. I don't know any EMH theory that alleges that efficient markets can predict the future with 100% accuracy, just that you can't beat them over time.

Overall, I'm not sure I get Davidson's point. He alleges that because no one can tell the future with certainty, risk managers should avoid all illiquid assets, and all assets that don't have a market maker with the resources to support the market.

I'm not sure I get it. First, there are illiquid assets in the world that are likely to have value. (Optioning a movie script, for example, or prospecting for gold.) Is Davidson saying that no one should invest in those at any price because we don't know if our assessment of the present value is accurate? Second, even in a market with a fantastically capitalized market maker backed by a central bank, there is still an unknown risk of catastrophic collapse. It might be smaller than in an illiquid market, but it's not zero.

Andrew M writes:

This reminds me of how common it is, usually in political discussion (surprise!), to forget that basic assumptions of mainstream neoclassic econ (perfect information, perfect competition, etc) constantly are not even considered, rather just assumed away, yet in so many of those cases are these assumptions verifiably untrue. Arguments hinge on these assumptions holding, but because academic econ cannot really produce models consistently without them, we are left utterly numb to evoking them while bridging the gap from academic exercises to policy.

Dan Carroll writes:

On the opportunity cost survey, what was the average beer consumption of the economists before they answered the question? I got the answer right, but if the reader was only reading it casually or not taking it seriously, then the knee jerk reaction would be to answer it incorrectly.

Tracy W writes:

Andrew M: okay, if these assumptions are verifiably untrue, and they're not even considered in political discussion, what's the problem?

I mean, I can get complaining that political discussion keeps considering false assumptions. I can get complaining that political discussion doesn't consider relevant true facts. But I don't get complaining that political discussion ignores false assumptions, isn't that what we want?

Scott Sumner writes:

I agree with most of the comments, although I think many are being too tough on economists. The field is very complex and those with different perspectives can look stupid if you don't know where they are coming from. For instance, I don't think raising the minimum wage is a good idea, but I wouldn't claim that all economists who favor it don't understand S&D.

Michael, The amount the person would be willing to pay to see Clapton has no bearing on the op. cost of seeing him. The op. cost describes what you give up by seeing him, not what you get.

Unlearningecon, No, he doesn't even come close to describing the formal version. The formal version says financial crises are unforecastable, and that people often cannot see the future. He says the opposite.


Jim Glass writes:

Anyone who doubts that economics is difficult for college students, many of whom are self-described econ students/majors/grads, just look at all the argument over the simple "opportunity cost" question that this post has inspired over at reddit:

http://www.reddit.com/r/Economics/comments/205gvs/economics_must_be_harder_than_it_looks_scott/

Enjoy.


Richard Besserer writes:

I have far more patience with Krugman than with Paul Davidson. Krugman is not the most diplomatic writer, but he's often at least interestingly wrong. For example, he knows enough about rational expectations in particular and how things are done in mainstream macro these days in general to describe their limitations intelligently.

Davidson clearly never learned rational expectations or modern modeling techniques well enough to have anything to say about it beyond "I don't understand it and it isn't compatible with my political prejudices, so it must be wrong." He's been giving versions of this speech for years (many of the papers on his website are recycled versions thereof), with few changes except in the buzzwords.

Comments for this entry have been closed
Return to top