WSJ: The Fourth Major Era of Computing Kicks In

http://online.wsj.com/articles/andy-kessler-the-fourth-major-era-of-computing-kicks-in-1412636637

Early in my career on Wall Street in the mid-1980s, a well-known strategist yelled at me for recommending a tech stock. “What are you stupid?” he said. “Don’t you know we are in the last year of a four-year economic cycle?” I didn’t know, and luckily I ignored him. The economic cycle lasted almost eight years, only to be briefly interrupted in 1990 before beginning another 10-year expansion. The old Wall Street adage “follow the cycles” may still ring true, but the cycles have changed.

Here’s how it used to work: From 1945 to 1982, the average business cycle lasted about 44 months. Lower rates at the Federal Reserve would at first stimulate demand but then ignite inflation, which would cause retailers to build inventory until rates rose to kill inflation. Manufacturers would lay off workers and close factories during the inevitable recession until inventories were depleted. Then the Keynesian pump and dump would start all over again. Those days are gone. First mainframe computers and then personal computers automated the pipeline of goods. New software and better supply-chain management smoothed out the entire process, ending the traditional inventory cycles of goods. Wal-Mart doesn’t order from Procter & Gamble anymore. P&G has a real-time feed of Wal-Mart sales, and supplies the right products at the right time. And you rarely hear of inventory problems at Amazon.

The result has been longer economic cycles, interrupted by big events. It was Saddam Hussein invading Kuwait in 1990 and 9/11 in 2001, both of which damaged global commerce. The 2008-09 recession, on the other hand, was self-inflicted: Too-low rates and exotic financial instruments created an excess inventory of houses and condos. Even that might have been avoided if a database of housing and real-time pricing had been analyzed like supply chains are now. Easy to do, as mobile and cloud innovation have lowered computing costs by a factor of 10 since then.

Which brings us to today.