Monday, March 14, 2011

Nuclear vs. financial disasters

Over the weekend, I have been pondering the potential problems at the two Japanese nuclear plants after the 8.9 earthquake on Friday. In the grand scheme of things, any problem seems to be contained and damage is likely to be relatively minor - though the Japanese are by no means out of the woods.

Dr. Josef Oehman, research scientist at MIT, in a blog post comment, believes that "there was and will *not* be any significant release of radioactivity", largely because of the design philosophy:
When designing a nuclear power plant, engineers follow a philosophy called “Defense of Depth”. That means that you first build everything to withstand the worst catastrophe you can imagine, and then design the plant in such a way that it can still handle one system failure (that you thought could never happen) after the other. A tsunami taking out all backup power in one swift strike is such a scenario. The last line of defense is putting everything into the third containment (see above), that will keep everything, whatever the mess, control rods in our out, core molten or not, inside the reactor.
The whole comment is excellent, concise in explaining the fundamentals of the plant's design and what happened. It is well worth reading. Oehman went on to give kudos to Japanese engineering:
The earthquake that hit Japan was 5 times more powerful than the worst earthquake the nuclear power plant was built for (the Richter scale works logarithmically; the difference between the 8.2 that the plants were built for and the 8.9 that happened is 5 times, not 0.7). So the first hooray for Japanese engineering, everything held up.
The Japanese been expecting the "Big One" for decades so this earthquake was hardly a surprise. They have built much of their infrastructure with a major earthquake in mind. It is therefore a testament to their preparedness that a far more powerful earthquake (8.9) struck Japan and left it with a death toll in the tens of thousands (and much of that was from the tsunami and not the earthquake). By comparison, the 2010 Haiti earthquake was far less powerful (7.0) but produced a death toll of over 300,000.


What about financial earthquakes?
By contrast, the world went through a financial earthquake in 2008 when Bear Stearns and Lehman Brothers collapsed. So it was a shock when I saw Bloomberg report last week that an "International Monetary Fund report shows that regulators haven’t gone far enough in taming potential financial-market excesses since the economic crisis began."

The IMF report stated: "We are at the moment even less well-prepared than when the crisis erupted in 2007." [emphasis mine]

There are macro risks everywhere. Should another financial earthquake occur, it should not be a surprise to anyone.

Have we learned nothing? Where is the redundancy that is found in nuclear power plants? Where is the "defense of depth", or any defense?

2 comments:

Sion said...

I am an engineer and as long as I have been following financial markets I have been wondering this. Except that I know the answer. Banks! They issue currency. They pay them-selves bonuses. Why worry when I'll be all-right is their attitude.

In addition to that there is a failure to understand the connection between energy and currency. Currency is a call on future labor and equivalent in some ways to stored energy. This has implications in regards to correct asset allocation.

And of course a la Steve Keen and his roving cavaliers of credit and the fact that neoclassical economics does not properly account for debt in its models.

Nick Gogerty said...

The nature of financial markets means redundancies are hard to design in as they usually involve a diversity of features or components in the design.

In fiscal environments the way to design robustness and resiliency into the system is to have loosely coupled (ie. less inter related components). These loosely coupled networks by definition have buffers in them.

One thing working against this is the drive for efficiency and ROC in finance. A tightly coupled system is by definition efficient.

For thoses interested in the topic. I teach a free course in Greenwich on Black Swans and systems design thinking. Join the black swan group on linked in. topics include, normal accidents, risk homeostasis, loosely coupled design, cascading failures, critical instability etc.