Sunday, March 20, 2011

For want of a nail, etc.

Some interesting articles in the NYT today on the economics of the earthquake/tsunami in Japan:

Stress test for the global supply chain, by Steve Lohr (3/19/2011)
  • "Chain" is misleading, because it really is a network, or a system ("Modern global supply chains, experts say, mirror complex biological systems like the human body in many ways"). Where there is redundancy and resilience, the disruptions are manageable.
  • "The good news for the world’s manufacturing economy is that the sectors where Japan plays a vital role are fairly mature, global industries." Mature networks have developed redundancy and resilience, and moved away from single source.
  • "Still, Japan produces a far higher share of certain important chips like the lightweight flash memory used in smartphones and tablet computers." Even if Japan makes "only" 35% of the flash memory chips, that is still a big chunk of source.
  • "The field of buying and shipping supplies has been transformed in the last decade or two. Globalization and technology have been the driving forces. Manufacturing is outsourced around the world, with each component made in locations chosen for expertise and low costs. So today’s computer or smartphone is, figuratively, a United Nations assembly of parts." Attempting to separate "globalization" (as it is meant today) and "technology" as two distinct things is silly -- globalization is a function of new technologies; capitalism in the age of electronics.
  • "That means supply lines are longer and far more complex than in the past." Which makes them more tenuous and vulnerable. The term for comlex procurement networks and long supply lines is "thin strands."
  • "The ability to manage these complex networks, experts say, has become possible because of technology — Internet communications, RFID tags and sensors attached to valued parts, and sophisticated software for tracking and orchestrating the flow of goods worldwide." Exactly -- this is why one can say globalization is a function of technology.
  • "'In the past, when you had a disruption, the response was regional,' says Timothy Carroll, vice president for global operations at I.B.M. 'Now, it’s globalized.'” So global supply networks should make it easier for producers to accommodate the Japan disaster.
  • Here is the really interesting insight for me: as the supply networks expand and complicate, "the difficulty and expense of seeing deeper into the supply chain increases." (emph. added) The problem becomes determining the impact on the parts used to make the parts, and the parts used to make the parts that make the parts, etc. etc., and even the raw materials to make the parts etc. etc. "For example, reports that a Mitsubishi Gas Chemical factory in Fukushima was damaged by the tsunami have fanned fears of a coming shortage of a resin — bismaleimide triazine, BT — used in the packaging for small computer chips in cellphones and other products."
  • "The Japan quake, some experts say, will prompt companies to re-evaluate risk in their supply chains." Rich, deep, redundant, resilient networks as risk minimizers (for more see a thing I did called "Networks and Globalization"). In this case, it might prompt "a shift from focusing on reducing inventories and costs, the just-in-time model, pioneered in Japan, to one that places greater emphasis on buffering risk — a just-in-case mentality."
  • Redundancy is one strategy -- multiple sources; resiliency is another -- the ability to quickly develop alternatives in case of emergency.
A Crisis That Markets Can't Grasp by Jeff Sommer (3/19/2011)
  • How do you price risks that are relatively remote, where there isn't much experience, if any? 9/11, Hurricane Katrina, the BP oil spill, and now the Japan earthquake/tsunami? "So perhaps a bigger question is whether the markets — in which we have come to place so much trust — can put a true price on outsize risks like this."
  • “Past performance is no guarantee of future success."
  • Nassim Taleb, who popularized the term "black swan" (referring to rare or difficult to predict high impact events) "argues that we have psychological biases that blind us to the enormous role played by rare events — like a 9.0-magnitude earthquake. And yet we rely on history for guidance."
  • "Whatever else they may be, markets are immensely complex counting machines. They assign values to products, whether computer chips or potato chips, based on the canny appraisals and gut beliefs of people around the planet. From day to day, it is often hard, if not impossible, to know exactly why a certain price moved the way it did. Those daily movements are often just the white noise of global capitalism."
  • And then "there are stretches of extreme volatility, periods when no one quite seems to agree about where anything is going."
  • And nuclear power... "The financial markets have always had a difficult relationship with nuclear power, largely because the costs — and potential risks — associated with nuclear plants are so huge. Day to day, nuclear power is cheap. But it is unclear how to accurately assess the cost of disposing of nuclear waste over the long run — or, perhaps, the cost of disasters like the one in Japan." Economists, politicians, capitalists may just ignore the externals, the environmental costs, either because someone else (namely you and me, the general public) will pick up the tab, or because it is too complex to calculate, or there isn't enough information to figure it out. "Professor Stavins said, 'You can argue that the markets have never really had a chance to price nuclear power.'”
  • "In the face of black swans — also known as fat-tail events, for the way their occurrences are distributed along a probability curve [those outlier events, low probability, but possible, like winning the lottery) — market pricing may be impossible."
The last comments relate to this final article:

Lessons from Chernobyl for Japan by Ellen Barry (3/19/2011)
  • The contaminated area around Chernobyl will remain contaminated for more than 300 years.
  • "One had to look at Ukraine to understand the sheer tedium and exhaustion of dealing with the aftermath of a meltdown. It is a problem that does not exist on a human time frame."
  • An eerie instance of an abandoned civilization: "The wild world is gradually pressing its way in... 'This is a city that has been captured by wilderness,' he [ said. 'I think in 20 years it will be one big forest.'"
  • According to one worker involved in maintaining the shell around the defunct reactor: “'Nobody knows what to do with what is inside,' he [Anton Yukhimenko, who leads tours of the dead zone] said. 'There will be enough work for my children and my grandchildren.'"


Sunday, March 06, 2011

Computers and chess

I continue to be fascinated by the ways in which computers (and new technologies in general) change the terrain in so many (all?) fields. Dylan Loeb McClain's chess column in the March 5, 2011 New York Times, titled "Still No. 1 in the World, by the Thinnest of Margins", comments on the fact that the world's three top-ranked players are only separated by a range of 9 points (out of 2,800 points). Garry Kasparov, the top-ranked player for 20 years until he retired in 2005, was, McClain notes, was often 30 or 40 ratings points ahead of his closest rival.

What I find of interest is that McClain attributes this to the fact that players today have access to the same databases of games (want to review the Immortal Game? you can find it in any number of places, and step through each amazing move. What about a match between Salo Flohr and Max Euwe, 1932? Yep. Or all of Bobby Fischer's games, or ... see, or Plus they have access to powerful computer chess engines to experiment with. And for the average punter, online play can dramatically expand one's board experience. So the tools of preparation are more evenly available: "Today, players have access to the same databases and computers when they train, and native ability is more important than ever" (which is another interesting insight).

Also see this post re: an article by Kasparov on computers and chess.