It is time to stop quivering in our boots in pointless fear of the future and just roll up our sleeves and build it.
- Ray Pierrehumbert

Friday, October 24, 2008

Failure of Networked Systems

This article, "The Failure of Networked Systems", on The Oil Drum explains matters very well.

It is sort of an engineer's twist on the tradeoff between risk and return. Go read it.

It applies both to the immediate financial mess and the even bigger picture reservoirs-and-flows views of the earth system.

What I have to add is this:

The current financial disaster is based on people deluding themselves that they had eliminated risk, when in fact they had coupled risk. The consequence is that small failures were avoided at the expense of big failures.

The whole setup of modern human activity makes a comparable error. There is no such thing as unlimited growth. All growing systems reach limits. The most casual understanding of exponential growth (h/t HR) makes this clear.

Either fuel supply or carbon waste are likely candidates to be the limit we hit first, but there are others. It doesn't matter. The "growth forever" idea is really "growth until it stops". If we base everything we do and everything we think on an assumption of growth, we start to build things in to protect the growth.

Much of government of the past century has been about protecting the growth. Sooner or later it is doomed to fail.

Has this just happened? Has the system reached old-fashioned bubble-popping so emphatically and so hard upon its physical limits that we will be unable to right it? Maybe, but probably not.

The problem is that righting it is not what we need to do. What we need to do is relax.

What we will inevitably try to to is rebuild the tightly coupled growth-dependent system that has spectacular failures built into its whole M.O. Realistically, some of this is unavoidable at this stage, but it's an ill-timed distraction. What we ought to do, instead, is reduce growth dependency and increase redundancy and resilience.

We need to convert to a world where less wealth gets created, and less wealth gets destroyed.

This is the relaxation scenario; it is easier on everybody, but it will take some creativity. In a perceived crisis, can we find the creativity to say, "no, we don't particularly want things to get back to normal"?

Resilience, not growth, is the goal of our time. We need to build a world where time to think and time to enjoy and time to care is valued more, where time to achieve and money to spend is valued less. Say you don't want no diamond rings and I'll be satisfied. Tell me that you want the kind of things that money just can't buy.

Trying to find sustainability in conventional economics at a time of stress is a category error if ever there was one. This mistake has its precedents and they are not encouraging.


Belette said...

"based on people deluding themselves that they had eliminated risk": not really; but they did think they had correctly priced it in. They were wrong about that.

David B. Benson said...

I have no idea how to correctly price a six sigma event. I doubt anybody does.

Michael Tobis said...

David, sure, but this financial event was far from unlikely. This seems clear enough to me in hindsight that it should have been clear in foresight to people who were actually playing the game.

Indeed, Greenspan is saying exactly that. He would have expected people to behave like grownups and not put all of capitalism at risk. That is exactly the mistake he is owning up to.

What aspect was anything like six sigma here?

There's nothing stochastic about the current fiasco.

As I have come to understand it the credit default swap ponzi scheme was going to blow up eventually and the only question was when. I see nothing unlikely about it at all.

David B. Benson said...

The people playing the game might have only been in it for their high salaries and bonuses. They might not have thought about the consequences at all.

Of course, their are always prophets proclaiming the end is nigh! I suppose sometimes one ought to listen to them.

In today's TNYT the article about Greenspan stated that he said words to the effect that the only thing he would change is having the derivative makers required to keep a portion of the 'paper' they created. Would that have actually been enough?

Anyway, I amsure we agree that unregulated capitalism is horrible. Dark, satanic factories and all that.

Michael Tobis said...

Ah, I see the "6 sigma" is a reference to the first comment on the Oil Drum story, and another very interesting piece.

I suppose there's a way of looking at it as 6 sigma if you just look at the behavior of a line. This is the statistician's worldview in a nutshell, and it cause some confusion in climatology as well.

The fact is that we know, or at least should know, more than nothing.

A huge increase in building construction simultaneous with a huge increase in building prices really ought to have told all of us that something odd was going on. But I forgive us amateurs more than the people who claime to be experts.

David B. Benson said...

Greenspan, Cox tell Congress that bad data hurt Wall Street's computer models

Another case of unduely trusting results because those results came out of a computer.

Bah, humbug.

Dano said...


bad data hurt economic decision-making. Bad intelligence hurt military decision-making. Bad science hurts...

We get the idea.



David B. Benson said...

Here is David Corn's take on Greenspan's testimony:

Alan Shrugged

Hank Roberts said...

Dang. And when every location on Earth depends on having some essential like wheat or soy shipped in from a long way away, and shipping suddenly declines for economic reasons, we are suddenly in a condition where the local population far exceeds the carrying capacity available.

This is an index for shipping (oceangoing dry freight):

As predicted:

" in Pane D, 'carrying capacity' has been represented by two different curves. A major fraction of the recent, apparently high carrying capacity for human high-energy living must be attributed to temporary resources —i.e., non-renewable fossil acreage, the earth's savings deposits. In Panel D, it is optimistically assumed that the component of carrying capacity based on renewable resources has remained stable so far. But it is recognized that serious overshoot, induced by temporarily high composite carrying capacity, will at least temporarily undermine even the sustainable component. "Energy plantations" for example (one of the Cargoist3 proposal's), will tend to aggravate the competitive relation between our fuel-burning prosthetic machinery and ourselves; land taken over to feed technology will not feed humans. So "temporary carrying capacity" is shown actually dipping below the horizontal line for a while, before it recovers and becomes again simply "carrying capacity". The lesson from Panel D is that crash caused by the exhaustion of phantom carrying capacity by Homo Colossus could preclude a later cycle of regrowth."
The Ecological Basis of Revolutionary Change
Author: William R. Catton, Jr.
Foreword by Stewart Udall

$27.00 Pub Date: 1982

Interview with Catton, August 2008:

How long does it take to get the food supply moving again?

David B. Benson said...

Failure of rating agencies:

Credit Rating Exec: "We Sold Our Souls to the Devil"

The conclusion I draw is that neither an only company-pays nor an only investor-pays approach will work.