January 2, 2009

Evidence of Over-Optimization

More evidence of over-optimization—explained here by Nassim Nicholas Taleb and Benoit Mandelbrot—of the world economy: Toyota takes a step back from just-in-time production.
Toyota Motor Corp. and Honda Motor Co., Japan’s two largest carmakers, may modify their so-called “just-in-time” manufacturing system to avoid possible supplier bankruptcies disrupting production.


Plunging demand in the U.S., the world’s biggest auto market, contributed to Toyota on Dec. 22 forecasting its first operating loss since 1938. That was the same year the carmaker fully adopted the “just-in-time” model, according to its Web site. Under the system, companies avoid stocking inventories, preferring to take delivery of components as they are needed, to cut expenses.

Any emergency measure would be costly, analysts say. Increasing stockpiles would mean renting warehouse space to store parts and supplementing components from overseas would increase shipping costs.

“We’re considering many scenarios for possible outcomes” from a U.S. automaker’s collapse, said Yasuko Matsuura, a Tokyo- based spokeswoman at Honda Motor Co., Japan’s second-largest carmaker.

Measures may include increasing inventories and doubling sources to buy parts, Matsuura said. “We also have strength in having global models such as the Accord and Civic, so we can share parts from other regions for those models.”
Mandelbrot and Taleb have a solid theoretical background for speaking about over-optimization. If you have the time, you should check the whole episode of Online NewsHour. Here is a short excerpt:
PAUL SOLMAN: So, getting back to your fundamental work and insight, this is a system that can become turbulent or is inherently turbulent, that doesn't have enough of a buffer, and that's the danger?

BENOIT MANDELBROT: That is not well-understood. In fact, that is misunderstood for which tools have been developed which assume that changes are always very small.

If one of them comes, nothing bad happens. If several of them come together, very bad things have happened. And the theory does not take account of that, and the theory doesn't take account of very large and sudden changes in anything.

The theory thinks that things move slowly, gradually, and can be corrected as they change, whereas, in fact, they may change extremely brutally.

NASSIM NICHOLAS TALEB: Now you understand why I'm worried. I hope I'm wrong. I wake up every morning -- actually, I don't wake up every morning now. I start to wake up at night the last couple of weeks hoping that I'm wrong, begging to be wrong.

I think that we may be experiencing something that is vastly worse than we think it is.

PAUL SOLMAN: And we think it's pretty bad.

NASSIM NICHOLAS TALEB: It's worse. Of all the books you read on globalization, they talk about efficiency, all that stuff. They don't get the point. The network effect of that globalization, OK, means that a shock in the system can have much larger consequences.

No comments: