Saturday, November 29, 2008

The overoptimized society

For former Wall Street hedge fund manager and self-styled student of uncertainty Nassim Nicholas Taleb an important cause of the current financial meltdown is best described by ecological science. The system has become overoptimized. The consolidation of finance into the hands of fewer and fewer large players--banks, insurance companies, investment banks, and giant hedge funds--has made it less vulnerable to frequent crises, but more likely to produce a severe crisis when there is a breakdown in the system.

What used to be country-specific or regional crises, now become worldwide crises. In the past we've had the Mexican crisis, the Asian crisis, the Argentinian meltdown and most recently the utterly devastating hyperinflation in Zimbabwe. But none of these became global crises.

"It's vastly more optimal to have one large bank than 10 small banks. It's more efficient," Taleb told The News Hour with Jim Lehrer recently. "[But,] when one bank, [a] large bank makes a mistake, OK, it's 10 times worse than a small bank making a mistake." The moral of the story: A world with a lot of small banks is far more resilient than one with a few large banks. That's the kind of result one would expect in biological communities, and it turns out to be true, not surprisingly, in human communities as well.

But overoptimization isn't just limited to the banking industry. In fact, it is everywhere, and it makes for vulnerabilities across multiple fronts that quite often interact with one another. We've built a system too complex for any human to understand. Therefore, when something major goes wrong, no one can be sure how to correct it.

Witness the floundering attempts to revive the comatose credit markets. The seemingly incoherent policy shifts exhibited by U. S. Treasury Secretary Henry Paulson and Federal Reserve Chairman Ben Bernanke are not so much a result of incompetence as a reaction to the opacity of the global economy and the inability of anyone to grasp its workings or interpret its supposed signals.

As for the vulnerabilities across multiple fronts, one need look no further than the world's ports. The financial crisis has slowed many to a crawl as exporters worry that importers on the other side of the ocean may not be able to pay them. Banks are reluctant to issue letters of credit guaranteeing payments when they can't be sure the bank on the other side of the transaction is sound. This has driven dry cargo rates down 90 percent from their highs this year forcing some shipping companies to simply idle freighters. In addition, several shipping companies are on the verge of bankruptcy. And, that means orders for new ships are plummeting as well driving shipbuilders toward bankruptcy or at least consolidation.

The tight coupling and vast size of the globe's major economic actors, once hailed a triumph for economic efficiency, has now become an Achilles' heel. Only a few months ago the globe's just-in-time delivery system was strained beyond the limit. Now, as demand has fallen off a cliff, it's spewing out so much stuff that the Port of Long Beach doesn't know where to put all the imported cars coming ashore. Feast or famine, the system wasn't designed for surges or sudden drops in demand.

Overoptimization has taken place in agriculture as well. We are now hostage to such potential crises as the so far unstoppable wheat rust which threatens to knock down wheat yields worldwide. Farms that grow and rotate multiple crops with multiple varieties would be less efficient, but far more resilient in the face of such plant diseases.

And, with the huge, but clunky U. S.-Canadian electrical grid, we are subject to large consequences from small disturbances. Overgrown tree limbs in Ohio were said to blame for the 2003 power outage that deprived 50 million people of electricity from Detroit to New York and from the Ohio River to northern Ontario. A distributed power system made up of a large number of individual home and business generators might be less efficient; but it would be much more resilient. A failure of even many parts of such a power system would leave most of the rest functioning.

The effects of the overoptimized society have been seen in price swings as well. Small changes in the supply and demand for oil can send the price soaring or plunging, and we have seen both this year. The same has been true in the grain and base metals markets.

What all this tells us is that the pablum we've been fed about the merits of globalization has masked its dreadful vulnerabilities. We have created a system that nobody understands and nobody can fix when it falters. We were told that the global marketplace could heal itself and correct imbalances. Well, obviously it can't, not without mowing down an awful lot of people who did nothing to cause the current financial meltdown.

That's the problem with a complex, tightly interconnected network. When it spirals out of control, it tends to create a cascade of problems everywhere. Nassim Taleb says he used to get up in the morning and worry about what will happen to our out-of-control financial system. Now, he wakes up in the middle of the night and wonders how bad it could get.

Maybe, just maybe, it's time to redesign our society with resilience in mind. If things keep going in the current direction, we will surely be forced to.


Frank Gifford said...

Thanks Kurt for your insights!

It seems clear to me the current system is broken beyond repair. This is self evident in the bailout methodologies. Consolidation wrought increasing fragility, and the nature of the market seems to select for undesirable human characteristics like greed, lying, cheating, ... However, when watching a YouTube of some mainstream media talking heads discussing the financial crisis, there was a discernable low level hysteria manifest by all the speakers. By and large, Americans seem comfortable with their lives and loathe significant change. After all, "a familiar hell is better than an unfamiliar heaven". So how do we redesign for resilience? According to the Native world view, we first must visualize the change we wish to see. When we change ourselves, we change the world, according to Ghandi. We seek to model some of these methodologies at our project EntropyPawsed.

Anonymous said...

The global supply chain risk has additional issues: when it was still full there was no inventory in any of the US-side nodes. Tax law and profit hunger have created this monster. It's blind dogma that says you shouldn't have any inventory for anything anymore.

We actually saw how close to the edge supply chains are with Katrina when the chains were severed and suddenly there's no food or fuel. Even people in the oil business see the risk with energy.

Supply chains and JIT have been optimized in so well that any time in the last 5-10 years if you wanted something not on the standard menu, you'd suddenly see the real transit time for goods through the chain which ranged from 2-18 months for even minor things like odd-valued resistors or screws.

JIT meant that these had to recursively trigger every upstream supplier up to raw materials in order to custom manufacture a run because no one in the chain isn't doing JIT now. This is the problem with this kind of "innovation" - once everyone is doing it, the advantage disappears and everyone faces the downsides together, namely, that essentially we've become as inflexible as any Soviet planned economy.

Add to this the depth of the US-side of any supply chain of any business only averages 2-3 hops and its overseas. Far less is manufactured in the US than most people who know we've outsourced everything even believe which means if we don't make it here and global supply chains break we are simply SOL for that item and any downstream consumer of it.

My background is in reliability engineering and I can assure you that what we have is dangerously unreliable. The analogy I use is that we've decided that all the redundant hydraulics and electrical systems of our economic 747 have been judged wasteful so we tore them all out for the sake of profit ignoring the fact that eventually we'll have a situation that needs that redundancy just to stay aloft.

At the moment, I fear, we may have already run into that situation.

But more frightening yet are those people calling for "centralization" to fix food supply risk or security of global supply chains. This is as wrong an idea as putting out a fire with gasoline. Making things more centralized will only create worse single-point failures and reduce overall efficiency in terms of cost and availability at the same time.

Only adding redundant, independent and competing economic players can improve the risks, just like adding redundant, independent hydraulics in a jet improves reliability and security to its operation.

::athada:: said...

Sounds like Wendell Berry was right...

Anonymous said...

Though not exactly a case of over-optimization (arguably using more energy), the recent push toward a 'google-ization' of the desktop along with applications, driven by lust for complete market domination (good summary here) will result in a similarly brittle system. A local grid failure at a server farm (which can draw 100 megawatts these days!) would have the potential to take down everyone's and everyone's business desktop, apps, and content.

Kurt Cobb said...

Thanks to Anonymous for bringing us some truly excellent intelligence from inside the logistics chain. And, thanks to Marty for showing how even one of the most resilient systems we have, the Internet, might have the resilience taken out of it. Frank, you are unfortunately correct that most people prefer a "familiar hell" to an "unfamiliar heaven." But kudos for trying to create the actual alternatives here on Earth now.

Unknown said...

Regarding, "Feast or famine, the system wasn't designed for surges or sudden drops in demand"; actually only half true. The system was specifically designed not to accomodate surges or sudden drops, as both of these conditions require cushions, and cushions in a "just-in-time" paradigm are by definition inefficient. In other words, the risk of not having a cushion has been undervalued, while the costs of maintain said cushions have been overemphasized.

Regarding, "We were told that the global marketplace could heal itself and correct imbalances. Well, obviously it can't, not without mowing down an awful lot of people who did nothing to cause the current financial meltdown"; this is the essence of the neo-classical economic view - that markets under unsustainable conditions will self-correct. Yes they will, but at what consequence to the participants in the market? Demand destruction = destruction of economic activity, as we may well be witnessing now with the 9% year over year decline in oil consumption attributed to the $150 oil this summer.

Unknown said...

Addendum to my first point above: this is a clear manifestation of the consequences of the current tendency to value a short term view more highly than a long term view. In the short term, unforeseen events are more unlikely than in a long term view.

Anonymous said...

love your blog good stuff