Bryan Caplan  

When You Put It That Way, Scott...

Crisis Dialogue... Scott Sumner's False Dichotomy...
I can't believe how much excellent material Scott Sumner hides "below the fold."  A prime example

Whenever I read opinion pieces by almost any macroeconomist-- Keynesian, monetarist, new Classical, Austrian, etc, there is almost invariably a point where alarm bells go off.  At some point the economist will make an assertion that seems to me to be in conflict with the EMH.  And after that point I have trouble taking anything they say seriously.  I keep thinking "If you're so smart . . . "

If you read through all my posts you will have trouble finding a single assertion that is at variance with the EMH.  In contrast, pick up almost any other economist's take on the current crisis and you will almost invariable find at least one assertion that conflicts with the EMH.  It's always something like "the root cause of the crisis occurred years earlier when . . ."

1.  The Fed set interest rates too low

2.  Regulators let banks make excessively risky loans (or if you're a right winger-encouraged them to make risky loans.)

3.  Americans didn't save enough

And so on.  In contrast, I believe that the depression was caused by events that took place in September and October, when the markets actually crashed. Which depression?  All of them--1929, 1937, 2008, etc.  And as far as I know I am the only economist who believes that all of these depressions were caused by events that occurred in those two fateful months.
This argument is so persuasive that I'm strongly tempted to second Scott.  My main proviso: He's got to let me count the public's panic as an "event that occured in those two fateful months."  How about it, Scott?

Comments and Sharing

COMMENTS (21 to date)
Pedro P Romero writes:

By accepting this line of reasoning, one is forced to believe in the extreme fragility on which our credit institutions are based on. Policy mistakes during those days, plus Prof. Caplan's point about panic were enough to cause the recession. Did it not those two elements were also present before "those two fateful months" in 2008? How many "fateful" september and october between 1937 and 2008 have been? What about 9/11? What about 10/87?
I think a recession like the current one ;ie of global proportions; unfolded somehow in an accumulative fashion. Of course, it should not be blamed on 'original sins' like the foundation of the Fed.
It should be thought as a crisis that began with small sparks in certain nodes of the market network, that turned out to be critical (even though they might have been relatively small). Then it spread out slowly until it was too late to contain it. Like any other unintended consequence of policy mistakes.

Greg Ransom writes:

We're suppose to take the EMH seriously, or is this a comedy bit?

I can't tell.

Why do stock prices rise and fall? Well, because on each of those days, people wanted to pay more or less for them. Simple.

I don't like all of those other more complicated explanations. (smile)

8 writes:

It's equivalent to a drug user who does heroin, coke, meth, weed, crack, with some prescription pain killers mixed in, for 10 or 15 years. Then one night they OD and die.

EMH is efficient at showing us what the majority are thinking but it doesn't tell us if they are right or wrong. Trends can persist for a very, very long time and panics are realization events, not changes in reality.

I don't think its odd that people didn't get rich off of what happened because these ARE events. He's right that it happens over a very short period of time. Austrians and others make their predictions years in advance. A trader who profits from it must first know the story and anticipate the change. Then decide what is the best course of action, knowing that things often happen unexpectedly, and then bet on it. As a personal example, in August 2007 I told a friend I should short Bear Stearns and put the proceeds into Yen futures. I didn't do it, because I didn't want to take the risk. Once the event hits, the cost of entering these trades escalates and the profit window rapidly closes. And the reason such large profit exists is because the risk is so great. I can tell you the roulette wheel will hit 15 eventually, but I'm not going to risk betting on it.

Mike Griswold writes:

This critique seems misguided, unless we are quibbling about proximate versus distal causes. To say ex-post that decision x many years ago created an instability which ultimately exacerbated outcome y does not violate EMH. A market can be efficient only ex-ante, not ex-post. Large binary events (such as housing crash / not housing crash) will be efficiently discounted to some expectation in the middle. This does not change the fact that a realization of the negative outcome will change the ex-post value of the market dramatically lower. When you are efficiently discounting disaster and not-disaster simultaneously, the outcome will have a big effect on the future value. This doesn't mean that Schroedinger's cat wasn't killed by nuclear decay....and it wouldn't be remiss for an historian to point that out.

ThomasL writes:

Is this view of EMH self-canceling? I think it essentially precludes the possibility of large-scale mistakes. The market (working efficiently) prices the risky propositions accurately. Can you panic efficiently? Why panic if the assets are accurately priced, isn't that inefficient?

I would propose that:

1) People can make mistakes.
2) They may make them at any time.
3) The effects of those mistakes are not always immediately realized.

Applied to the present case, a mistake was made: either purchase into a MBS at $40/s or its sale at $2/s: one or perhaps both of those decisions was wrong. (Or substitue the actual loan transaction instead, either one works.)

If the purchase at $40 was a mistake, then it was made in 2006 while it wasn't realized until 2008; if sale was a mistake, it was made in 2008 but will only become known at some undetermined point in the future.

Isn't this simple exposition sufficient to point out that mistakes often have a time lag? Lag is evidence of inefficiency.

What caused people to make these mistakes is a worthy subject for investigation, but it doesn't really matter. If you first admit that mistakes can be made, those mistakes will be, broadly defined, inefficiencies.

If ones believe the market is efficient (but not perfectly efficient, as perfect efficiency precludes error), one will believe the causes to be very close in time to any results. If one believe it to be chaotic, the causes can stretch well back in time.

I'm more of the chaotic view.

Rimfax writes:

The Wikipedia article pretty much takes for granted that EMH is invalid. (Or am I misunderstanding?)

Not that I'm taking Wikiality too seriously, but if EMH is as discredited as the article suggests, then why is this Sumner's nerd gate? If it isn't as discredited as the article suggests, then where are it's advocates on one of the most important collaborative knowledge bases on the planet?

pushmedia1 writes:

"1) People can make mistakes.
2) They may make them at any time.
3) The effects of those mistakes are not always immediately realized."

To refute EMH you have to assume the above holds for ALL agents in the economy at once and that these errors are correlated. This makes your criticism from psychology seem implausible.

EMH doesn't preclude agents committing errors, it just supposes those errors cancel each other out.

Greg Ransom, if this is a comedy, you're very rich.

Pop Psychology - Investment Bubbles

From The Atlantic by Virginia Postrel 12/17/08

A rise in price above a reasonable value (a price bubble), followed by a fall back to that value (a crash), seems to be a part of psychology rather than bad mathematics or a lack of information.

In lab simulations, all of the participants knew the expected value of their investments, but there was some uncertainty along the way. Price bubbles and crashes occurred in 90% of the simulations. They had all of the information, and they still created bubbles.

Dr. T writes:

I know this is a site for economists, but I can usually figure out what's being discussed. But not this time. WTH (what the heck) is EMH? Economists Mimicking Hicks? Econometric Model from Hell?

Scott Sumner writes:

Thanks Bryan, As long as you define "panic" as "correctly ascertaining that the monetary authority was about to embark on a dramatically lower NGDP growth trajectory that would plunge the world into depression" then I am completely with you.

I do agree with the chaos theory discussed above in one sense--the deeper causes of the recession occurred much earlier, But we know from the lofty levels of the stock (and commodity) market as recently as May 2008 (only modestly below record levels) that the current disaster was not viewed as inevitable. I don't doubt that had the sub-prime crisis not occurred the Fed would not have screwed up monetary policy. I understand that. But the sub-prime crisis did not make that screw-up inevitable. That's why markets crashed last fall, not during the previous 12 months when we knew we had a big problem with sub-prime loans. When you go to a dramatically lower NGDP growth trajectory it makes the debt crisis far, far worse. Not just this time, but every time NGDP falls during a debt crisis. Why is nominal GDP important. Because people have nominal debts. When their nominal income falls they default on nominal debts. Markets had been lulled by the Great Moderation into thinking the Fed could keep NGDP growing indefinitely. And they could have if they had adopted a NGDP futures targeting policy.
If this recent collapse in NGDP had been baked in the cake by policy errors earlier than last fall, then the markets would have crashed earlier than last fall.
BTW, I am talking about much more than just the stock market, but use that for convenience. The real estate sector got much worse last fall, commodities crashed, inflation futures plunged, the dollar soared, etc.

ThomasL writes:


I may have glossed over it a little, bu my basic thrust was towards a price error from misassessing the risk. An error in pricing is almost certain to mislead a lot of people at once. I would not claim that it would affect all agents at once, but it could affect quite a few, and the bigger the error the wider the sphere of influence, as it will draw people into the seemingly wonderful deal.

A very real counter-argument is that the risk was revealed by the high return. If these were truly low risk, they should have had low returns.

This is not a highly sophisticated answer, but I'd lie that to people's natural and well demonstrated ability to discount even obvious risks.

The "it just exploded" view more or less insists that people were acting very efficiently right up to September.

I will propose a different view conceding that:

1) The market, overall, tends towards efficiency.
2) When it is acting efficiently it is pretty stable.

However, I think it is possible for certain institutions, such as government, to introduce elements into the market which channel resources into one favored area at the expense of others, for no particular market reason. That distortion undermines the notion of a truly free or truly efficient market. People have a natural ability to discount risk, particularly when after trying something they experience no immediate adverse effects. Those distortions, combined with the downplaying of perceived risks, work to draw in further resources with erroneous pricing, and the whole system becomes subject to increasing positive feedback. Systems which are sensitive to large amounts of positive feedback are inherently unstable. At some point the curve gets too steep, there is one adjustment too many, and the feedback amplifies the tweak to an uncontrollable level.

I think this is a simple but basically more accurate view than EMH. If you have an efficient market operating efficiently it should only be subject to tremendous shocks closely associated in time with any effects. If, however, you’ve got a system which is inefficient (ie, distorted) and subject to positive feedback as a result of those inefficiencies, you only need a very small and possibly unrelated shock to upset it.

I don't think the inefficiencies are necessarily naturally occurring, I rather think they are unnatural, but they do exist in this system.

aretae writes:

Dr. T

EMH = Efficient Markets Hypothesis

johnleemk writes:

While I generally agree with the view that government policy can abruptly disrupt the workings of the market, I can't help but think that to some extent, markets (especially financial markets) are by nature prone to such abrupt disruptions like bubbles. After all, Tulipmania and the South Seas bubble arose in an era of hardly any government regulation. Modern markets have a lot more regulation, so it's not surprising that bubbles can be traced to government involvement, but that may just tell us that markets are prone to making mistakes on occasion. I am not convinced that government is the primary/only source of bubbles and the like.

Vangel writes:

I find it fascinating that after all that we have seen people can still believe in EMH as it is taught. The bottom line is that market prices are set by the marginal transactions that are taking place in a system that is not entirely stable and where changes in liquidity can cause massive swings in short periods of time without regard for true fundamentals or what is usually referred as intrinsic value. EMH is a joke taught by incompetent charlatans who are too removed from reality to have much in the way of common sense.

Greg Ransom writes:

Scott writes:

"If this recent collapse in NGDP had been baked in the cake by policy errors earlier than last fall, then the markets would have crashed earlier than last fall."

All I have to say is that there is no reason whatsoever to believe this, especially when we know how money and credit can distort the whole time structure of production across all of the various heterogeneous stages of production -- creating misallocations across that structure that only reveal themselves later in time.

Oh. I forgot. Chicago trained economist work with "models" that don't have different production processes.

Troy Camplin writes:

Is the first sign of something the same as the beginning of it? How much of an ice burg can actually be seen? A whale pops up to breathe, and we see its spout -- is that the true beginning of the whale? Isn't there a long snout in front of it? How often is an artist's first successful work his first actual work? I suppose Scott also thinks successful CEOs appear out of thin air as well?

Jesse writes:

If there are limits to arbitrage, then there is no reason for markets to be fully efficient.

baconbacon writes:

"But why do you need to predict the peak? Asset markets are incredibly volatile, why not simply buy when markets are more than 20% undervalued and sell when they are more than 20% overvalued?"

This is a clear sign that the author hasn't thought deeply about the issue. The harder it is to predict a peak the harder it is to predict correct entry points. Take the housing bubble for example- the Case-Shiller 10 city index is at 100 in 2000, lets call this 'properly valued'. It hits 120 in 2002 and you short the housing market considering this to mean the market is 20% overvalued. Bye Bye bankroll, the 10 city average ran up to 225 in 2006- and despite the burst were still above 150 in the beginning of 2009. So sorry, so sad, you have no money left to ride the massively profitable collapse because you got out irrationalized by the markets.

Stephen writes:

Why restrict the explanation to just those "two fateful months?" Why not just those couple of days, hours, minutes? EMH appears to be a decent hypothesis, given no government intervention is distorting prices. However, prices do get distorted and hence the available information in the market cannot be correct.

Dr. T writes:

Thanks, aretae, I'd seen the term before but not the acronym.

Comments for this entry have been closed
Return to top