08 June 2014

Piketty and Growth Rates

I'm working my way through Piketty (Capital in the 21st Century).  He is decidedly pessimistic about growth rates.  And I think he misses the mark.

He gives a table (page 94) showing historical per capita annual growth rates in output.  The basic pattern is:

DatesWorld Growth Rate

From this, he concludes that in the 21st Century, growth rates will be around 1.5%.  His basic thesis is that growth rates increase when a society is "catching up".  The high growth rate from 1950 to 1970 is then due to rebuilding after the second world war, with growth rates returning to he mean (especially in Europe and the Americas) after 1970.

However, per capita growth rates tend to correlate with a couple of things:  Energy consumption and education.  The slump in growth rates from 1970 to 1990 can be interpreted as the world having bottlenecked on oil production.  From 1974 to 1985, the real price of oil was between $60 and $100 per year.  Prior to 1974, the price was down at $20/year, and after 1985, the price stayed in the $30 to $40 price range until 2004.  (source: EIA)

We are now moving away from using mining technologies to obtain energy to using manufacturing technologies.  Solar Panels and Wind Turbines are primarily a manufacturing technology.  When using mining technologies, we rapidly deplete the minerals that can be recovered with a given level of technology.  We continually need to improve the technology to discover and extract new minerals, and this keeps the price of the commodity high and the supply low.

With manufacturing technologies, we follow a learning curve that is primarily related to economies of scale.  As we increase production, prices drop.  Across a broad range of products, prices tend to drop by a factor of two as production increases by a factor of ten.  On top of this, there is a secondary rate of improvement due to improved technology.  Manufacturing technologies are not as limited by the supply of basic commodities.  Manufacturing processes can be adjusted to use different, cheaper, more abundant materials.  And materials can frequently be recycled.

Piketty writes (page 94) "Gordon's analysis is based on a comparison of the various waves of innovation that have succeeded one another since the invention of the steam engine and introduction of electricity, and on the finding that the most recent waves -- including the revolution in information technology -- have a much lower growth potential than earlier waves, because they are less disruptive to modes of production and do less to improve productivity across the economy."

However, we are really just starting to see what the new information technologies can bring to the table.  The intelligent car that drives itself has the potential to let people work online for an additional two hours a day while they commute to and from work.  And they will essentially eliminate deaths from car accidents.

Information technology has been providing a continual stream of improvements that have been incorporated into other products:  cruise control, ABS braking, air bags, cruise control that automatically speeds up and slows down in traffic, blind spot sensors, seat and mirror position memory, ...  As computer performance/price periodically doubles, the amount of available intelligence that can be incorporated in a design cycle rapidly increases.  We've been incorporating fairly small chunks of intelligence.  It's taken awhile for the chunks to get big enough to be incorporated usefully.  But now the chunks are usefully big and getting bigger.

The same pattern applies to productivity.  We automated the simplest tasks with small chunks of computer power available:  ATMs replaced tellers.  But bigger chunks of intelligence let us automate much more complicated tasks.