September 2022

Sun Mon Tue Wed Thu Fri Sat
        1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30  
Blog powered by Typepad

« How the Washing Machine Liberates and Educates | Main | Walter Williams Interview »


Feed You can follow this conversation by subscribing to the comment feed for this post.

I'm not denying the wisdom in Hayek's speech but the events in Japan can quite easily be explained by a simple "incentives" story.

Concerns about the risk of damage from earthquakes to Japan's nuclear reactors is nothing new . But no one has the incentive to do anything about it.

The reactor manufacturer (in this case GE) is explicitly absolved from any liability from such events - this is the standard in the nuclear industry globally. Witness the ruckus when India decided to go against this global "consensus" . And the plant operator in Japan is also not liable for the loss from extreme catastrophic events.

As seems to be the case in too many sectors in our "capitalist" economy, the taxpayer pays for all tail events. So why are we surprised that no one seems to bother taking precautions and insuring against tail risks?

Very insightful post!

Hayek in his brilliance may not have fully appreciated that complexity transcends the study of economics and the social sciences, and is alive and well in geophysics and biology.

It is no coincidence that contemporary debates in the complex physical science disciplines procede along almost identical lines as the debate regarding micro vs macro in economics.

Complexity is the subject itself. Spontaneous order is essentially chaos theory lurking in economics. Complexity is a staple of reality itself, and is an overarching theme.

The study of rules and institutions that have yielded Nobel prizes for Buchanan and Coase, North, and Ostrom are essentially the continuation of Hayek's program of political economy, and have an uncanny similarity to the study of boundary value problems in non-linear dynamics.

In the earthquake example, we see a non-linear social system and a the non-linear physical system with feedback between the two.

Hayek seems to have come to his thinking about complex phenomena as a topic in sciences studies as a respected student of 1940s era theorizing about purposively or functionally organized mechanisms and biological entities (e.g. radar tracking and brain mechanism) -- he discusses this topic in his 1952 book, _The Sensory Order_.

Darwinian biology and others domains of biology were always Hayek's go-to example of essentially complex phenomena that could be explained only "in principle" -- the counter-example to Popper's "Received View" falsifiability criterion of "science" that even Popper conceded (see Popper's autobiography).

In the 1970s Hayek became aware of physics of non-linear phenomena -- but doesn't seem to have been much of a master of that field.

This was before the publication of such things as _Chaos_ by James Gleik, who popularized the conceptual issues involved in non-linear dynamic systems, and the math developed by such people as Mandelbrot.

Mark Stone has a terrific article in the philosophical literature -- which exactly fits Hayek's point -- demonstrating the defeat of the Laplacean model of precise linear predictability model of explanation even in the physical realm.

Sralla writes,

"Hayek in his brilliance may not have fully appreciated that complexity transcends the study of economics and the social sciences, and is alive and well in geophysics and biology."

Let me add that the sensitivity to the measurement of non-linear systems -- yet the fact that they display systematic patterns -- is part of the grand collapse and defeat of the "Received View" of theories and science that dates back to Mill (or traced back even to Hume and before), which holds that science is about "laws" from which particular predictions in particular places in space and time can be deduced.

You still see people attempting to fit all of science in this debunked model (which also fell among other reasons due to internal anomalies discovered by its own champions).

The "Received View" of theories is exemplified by the work of Hempel, Popper, and Nagel.

The case for what might be called the "pattern" view of science can be found in various forms in the work of Hanson, Hayek, Toulmin, Wright, Kuhn, and others.

The view owes something to the Gestalt psychologists and to Wittgenstein, among others.


I have it from primary sources that Hayek contacted the physicists at Brussels (Prigogine group) and Stuttgart (Haken group) after the publication of his 1967 essay. He was treated with amusement and some disdain in Brussels, but was well received in Stuttgart, despite the weakness of his background in math and physics. This was partly due to the fact that Hermann Haken was a refugee from East Germany.

Regarding the matter of the the Japanese events, this is a complicated matter. 1) The probability distribution of earthquakes is well known, and while this one was an extreme event, it does not per se violate the distribution. 2) The problem arose from its location in the near offshore zone off Japan, thus triggering this enormous tsunami that apparently caused the majority of deaths that have occurred. 3) Japanese building codes have long been extremely strict in efforts to avoid earthquake damage. But they could not resist tsunamis of that magnitude. 4) While there have been huge amounts of hysterical publicity about the nuclear reactors situation, in fact only one has been the source of serious problems, Fukushima Daiichi #3, which is apparently now being gotten under control. As at Three Mile Island, it may well end up that zero people die as a result of the difficulties at the reactors, but the hysterical publicity arising from this will probably shut down expansion of nuclear power in many parts of the world (including the US, but maybe not India or China) for many years to come.

On the matter of Japanese earthquakes and the building codes, I was in a very bad earthquake in Tokyo in 2005. I was lecturing on speculative bubbles and said the word "crash" just as the quake hit. There was substantial damage and some injuries (mostly from falling bookcases and shelves with dishes), but nobody died. The strict codes date back to the great Kanto earthquake of 1923 when indeed many people died.

Yes, Hayek cited Prigogine in his 1978 essay, the afterword to _Law, Leg. & Liberty, Vol. 3, and I believe he cites him elsewhere.

I can't remember if there isn't correspondence between Prigonine and Hayek in the Hayek archive.

If I can ask, what is your source?

Barkley writes,

"I have it from primary sources that Hayek contacted the physicists at Brussels (Prigogine group) and Stuttgart (Haken group) after the publication of his 1967 essay. He was treated with amusement and some disdain in Brussels, but was well received in Stuttgart, despite the weakness of his background in math and physics. This was partly due to the fact that Hermann Haken was a refugee from East Germany."

Mandelbrot had an appointment at the research lab my dad once administered, but I never met the man. Sorry I didn't.

If I remember correctly, I think Hayek met Prigogine at some conference of Nobel winners organized by the Pope in the early.

When you have an Nobel in your hand, people tend to stop laughing.

organized by the Pope in 1980.

Wonder if Hayek learned of this work from Feyerabend or Watson.

"I have it from primary sources that Hayek contacted the physicists at Brussels (Prigogine group) and Stuttgart (Haken group) after the publication of his 1967 essay."


On the Brussels side I have it from Peter Allen, one of Prigogine's leading students. On the Stuttgart side I have it from Hermann Haken himself. I note that Stuttgart is quite close to Freiburg.

Any idea why?

"He was treated with amusement and some disdain in Brussels"

I am curious about why Crovitz chose to cite Hayek on this point when there so many other sources of this idea. Perhaps it was to show WSJ readers that it is okay for them to be sceptical about nuclear power -- even though they are pro-business conservatives.

Everyone has some agenda -- and that's okay as long as we know it.

Responsibility may be attenuated, but it is not absent. In a democracy, governments can fall if viewed as having mismanaged things over which they can exercise some control. Witness what happened in Iceland and then Ireland in the aftermath of the financial crisis. Also, in the UK. And to the Republicans here in 2006 and 2008 (war and the economy).

Japan is not alone in being struck by natural disasters of unprecendented magnitude. Witness the two big quakes in New Zealand, and also in Chile. And years of drought followed by a year of flooding in Australia.

How do you know the probability of something that has never happened? Hayek is relevant here, but so is Knight on uncertainty.

Eevn within the hard sciences, engineering is not just physics. There is an empirical element. No matter the planning, not all contigencies can be planned in advance. Experience brings up unexpected shocks, and designs are then improved.

"2) The problem arose from its location in the near offshore zone off Japan, thus triggering this enormous tsunami that apparently caused the majority of deaths that have occurred."

The general location of the earthquake was no surprise, and therefore it should not have been a major problem for the power plant. Open Google Earth and take a look at the sea floor off the northeast side of Japan. You don't need to be a geologist to understand what it is showing.

These types of earthquakes involving major subduction zones *often* produce large tsunamis. Ironically, the term tsunami is from a Japanese word meaning "harbor wave".

The fact that the backup pumps at the Fukushima power station were not designed to withstand a significant tsunami is a major engineering error that should not have occured.

Major convergent plate boundaries are known to generate very large earthquakes (even magnitude 9.0 quakes). The recent giant Indonesian quake and tsunami occured in an analogous tectonic setting.

It's not just predicting the possibility of an eathquake, but its magnitude. I believe the nuclear plants were built to withstand an 8.0 quake. They were built with the knowledge of their day, not our day.

The generators were built underground to proectect from cyclones. One can engineer for specific risks, but not all conceivable risks. There are tradeoffs.

The G.E. design was a bed design for the get go -- with well known un-necessary design weaknesses.


I think a combination of his weakness in math with ideology.


Geologists have known for over 40 years that Japan sits along a major compressive plate boundary/subduction zone. They also are well aware that very large earthquakes in excess of 9.0 are often (on a human timescale) generated in these settings.

The 1964 great Alaska earthquake measured a 9.2. It raised Kodiak Island some 30 feet, and caused a massive tsunami that produced damage all the way to Japan, and pushed a wave some 25-30 feet high inland in Southern Alaska.

The 2004 Indonesian quake was a magnitude 9.1, with the resulting tsunami killing over 230,000 people.

Go to Google Earth and look at the sea bed and tell me what these areas (Japan, Alaska, Indonesia) have in common. They are all along a subduction zone and oceanic trench that sits very close to shore. We know this is where mega-earthquakes occur.

Now if the reactors were designed for only an 8.0 and a moderate tsunami, it was engineering folly at its best.

In full disclosure, I'm a research geologist whose specialty area is in structure (faults and folds in the earth's crust) and tectonics, so if I sound too zealous about these sorts of things on an economics blog, please forgive.

Now on your point about tradeoffs, I fully agree and pointed this out during last years Gulf of Mexico blowout and oilspill.


You have the wisdom of hindsight. "Too zealous" are your words.

If there are multiple risks, against which do you act? Earthquake/tsunami, cyclone, etc? The solutions conflict.

If not nuclear, what energy sources should Japan use? As the Economist explains this week, among developed countries Japan has a singular problem in that regard.

More people have died in coal-mining accidents in China each year than died due to Chernobyl. Should China abandon its nuclear push in favor of an even more dangerous coal technology?

Japan, New Zealand, Chile all lie in the Circle of Fire -- very earthquake prone. Should they depopulate? Should New Zealand have abandoned Christ Church?

Nuclear does not = a flawed G.E. design over 45 years old.

It was an inferior design even for the time. There are far superior and safer cooling designs available.


I completely embrace the economics you are pointing out.

My point was much more remedial. It's good to use some common sense in engineering. As we have *learned* (Hayekian word of the day), it may not be a good idea to build a nuclear power plant whose operating life expectancy is 100 years (I'm not sure about this**) along the northeast coast of Japan 7 feet above sea level and design it for 10ft tsunami, especially when you do business in a country that invented the term tsunami.

It turned out to be bad business as the power company recently discovered.

K Sralla,

What was the bigger problem, the earthquake or the tsunami. My sense is that the plants were damaged by the earthquake, but the tsunami's effects produced the crisis.

Earthquakes aside, I've never understood why these plants have been constructed so close to large population centers. That is true on both US coasts. I'm visitng friends in Long Beach CA, and the San Ofre nuclear plant is 45 miles away on prime coastal land.

We have large expanses of desert in the US. Coal plants have been put out there. Why not nuclear?


I find it hard to believe these guy knew anything one way or the other about Hayek's mathematical competencies.

How they know anything about a scientist they certainly knew next to nothing about?

Barkley wrote,

"I think a combination of his weakness in math with ideology."

Jerry, nukes have needed a water source for cooling.

Theoretically, you could put a Thorium reactor in your back yard -- let's hope the technology makes it there.

If you fear radiation, stop eating bananas, stop flying in airplanes, and get a radon detector for your home.

The San Onofre plant does not use the flawed GE design.

I live about 15 miles away and don't think another thing of it.

Here's an article on the newer safer nike designs, using a passive cooling system:

Our knowledge make us able to predict that in certain areas great earthquake can occur; maybe our knowledge make us able to estimate their proper probability distribution, but we are not able yet to predict exactly when these event will occur, and this is for granted.
What can we do then? Get prepared.

Here we are falling into an "abuse of reason": s we know the event probability we can estimate that certain greatly distructive event can occur with a very low probability, maybe about once in 10.000 years. Our mistake is "weighting" the implied losses by this low probability, than get prepared only for its "expected value" (e.g. the complete world distruction can be pondered by such a low probability to reaching an expected value of some hundred dollars... but if the event occurs no dollar can ever be enough to fix the situation).

Here's the problem, the Value at Risk technique (VaR): a Richter-9 earthquake can occur only (say) 100.000 years, so it is not significant, hence we must prepare with implants able to cope with Richter-8.5 earthquakes which are far more frequent.
a 0.0001% probability means not that the quake will occur after 100.000 years, but that it will certainly happen WITHIN 100.000 years... maybe tomorrow.
When the event will exactly occur, is a missing information, thus abusing our reason by preparing for just "expected values" lead us to disaster from which we could not recover if we lived 100.000 years.

Add political incentives and so on, and you have the explanation why unimmaginable concatenations of disastrous events are so frequent.


A PDF based on historical data is not a very good way to look at this problem. We only have about 60 years of good data, and the historical records during the last several hundred years are not very good, as many of these areas were poorly inhabited. We know that events such as these tend to cluster with respect to time. There might be periods of several hundred years where major earthquake activity is fairly quiet, then another timespan where it is much more frequent. Since we do not know what regime we may reside in, it's hard to know if a probability function based on the last 50 years gives us much meaningful insight into future risks. This is a classic feature of highly non-linear systems. The "climate" of earthquakes is constantly moving.

Obviously, the frequency of an earthquake is inversely proportional to its magnitude.
However, for very large earthquakes, it should be stressed that the zone of high vulnerability to either ground tremor or tsunami extends out several hundred kilometers from the epicenter.

In the case of tsunami, the wave can radiate out thousands of kilometers.

So the chances of ever being at the epicenter are very small, but the chances of sustaining perceptable ground motion or coastal effects from a tsunami is much greater.

Since we have observed several mega-quakes in the last 50 years, and the population density and construction along the immediate coastal shoreline has greatly increased over this time span, engineering should probably take these events into account.


I never spoke with Prigogine, so do not know his view. I am guessing on Peter Allen, who seemed "bemused," saying to me that "we did not know what to make of him." So, just for the record, I cannot say that either the remark about ideology or math was the basis of this bemused attitude. You have it now more precisely. If anything, they knew too little about him at all. I was guessing from Allen's remarks and apparent attitude. In recent years Allen has been a prof of Management at Cranfield in UK.

Hermann Haken, who professed to like and admire Hayek, is less well known to readers here, as he has not received a Nobel. He was the longtime Director of the Stuttgart Institute of Theoretical Physics. He is known as the inventor of "synergetics" theory, which apparently Hayek became aware of and interested in, if not as much as the work of Prigogine and his group. It can be used to study large-scale phase transitions in systems, among other things, and one of the more interesting ideas is about how such large-scale phase transitions can come about in connection with a "revolt of the slaved variables."

Thanks Barkley.

As an undergrad I somehow ran into a book and some articles in political science on "synergetics" -- I think one or two of them made reference to Hayek.

The idea (and math) of phase transitions in systems is important.

I've found the study of the growth and collapsing of sandpiles, with sand pouring in from above to be very provocative.

The standard account is that such things are unpredictable in the sense of making simple linear predictions in time and size.

Hayek uses the metaphor of piling honey, built due to a constant pouring of the honey, as a metaphor for the effect of inflation on the structure of production.

I'm guessing you know the literature -- and the math -- of all this much better than I.

Hermann Haken has not written anything specifically on economics, but his close associate, Wolfgang Weidlich, has done so, as well as starting the field of "sociophysics," which has gotten a lot less attention than the related "econophysics." I published one paper by Weidlich in JEBO, whom I admire and like and think that he and Haken are underappreciated. Weidlich's top student out of Stuttgart is Dirk Helbing, now one of the world's top modelers of the nonlinear dynamics of transportation systems, who is based in Zurich and claims to be doing sociophysics.

Peter Allen has written some papers, almost all of them coauthored, in economics, mostly on the economics of technological change and earlier on urban and regional systems, with Prigogine and Stengers citing that work heavily in their widely-read _Order out of Chaos_. I published a paper coauthored by Allen with Hans-Peter Brunner on development policy in my 2009 Handbook of Complexity Research, published by Edward Elgar.

Barkley, you are filling up my Amazon wish list with these names! :-)

Dear Sralla,

you are right but that's not what I was saying.
VaR technique can be historically based but also parametrically based with "proper" forward looking perspectives.
What I meant, is that EVEN IF the probabilistic model would be right, we do not know when exactly and event will occur, as the probabilistic part of the model itself says that the extreme event WILL CERTAINLY occur within a certain lapse of time, and not precisely at the end of a period.

One problem is that, also because of normative prescriptions, various private or public entities MUST rely on Var; the other problem is that the same VaR technique is based on misconceived hypo's; all other problems are well know to Austrians.

Is anyone going to take a crack at Krugman's post on Hayek as a zombie?

The comments to this entry are closed.

Our Books