Monday, September 18, 2017

Hours Worked and Real GDP Growth

Over the long term, growth in real GDP correlates, naturally, with growth in hours worked.  Here, I have graphed the 10 year change in total nonfarm business sector hours worked (relative to population growth) and the 10 year change in real GDP.

There has recently been a decline in hours worked, which is partly demographic (aging baby boomers) and partly economic (the Great Recession).

There seems to be a long-term decline in RGDP growth of about 0.3% per decade.  Regressing real GDP against time and relative hours worked, we get the residual shown in the graph.  The regression suggests that for each additional percentage gain in hours worked, real GDP growth increases by about 0.68%.  I am using rolling 10-year growth rates.

By this estimate, growth in the 1990s was largely related to rising work hours.  Then, in the 2000s, work hour growth declined, but GDP growth continued to be normal.  Then, with the Great Recession, hours worked and GDP growth fell.  Now, growth in hours worked is low.  But, real GDP growth is low, even adjusted for slow work growth.  I suppose that could still be demographic, since older workers are exiting the labor force, and younger workers are entering.  It probably isn't a coincidence that the GDP growth residual declined as baby boomers were entering the labor force, rose for 30 years as they aged, and then started to decline as they began to leave the labor force, although clearly the sharp trend shift is related to the Great Recession.

tl;dr: There was a sharp drop in total hours worked beginning in about 2000.  This would be enough to drop real GDP growth rates significantly, and it probably did.  But, real GDP growth has been even lower than this sharp drop in hours worked would suggest.

PS: Broken record alert: For the past decade, residential investment as a percentage of GDP has been about 1% below long term averages and construction employment as a percentage of total employment has been about 1% below long term averages.  Add 10% to the 10 year measure for both hours worked and real GDP and your within range of normal.  It's about housing..... One might argue that there is competition for scarce resources and you can't just add real GDP growth without accounting for pulling those resources away from other uses.  But, I believe there has been a decade long search for the answer to the mystery of idle labor, long-duration unemployment, and a glut of savings that happens to coincide with this shift.  If ever there was little tradeoff for marginal investment, it has been now.


  1. Great post.

    A few years back I was fooling around with this series:

    As of now, total hours worked in the US are up slightly from 2007, and up about 5% from 2000.

    Well, if Kevin Erdmann is a broken record, then I am the repeating public announcement heard incessantly at some airports.

    It's property zoning, dudes. And property zoning (and other restrictions on housing supply) are ubiquitous enough to have macroeconomic consequences.

    The orthodox macroeconomics profession prefers to model and theorize in worlds with limited structural impediments. Then, the models work.

    But structural impediments can overwhelm theory.

    Real wages are not rising, probably falling, and housing costs play a huge role in that (especially in key cities). Supply and demand---there is less supply of labor when wages are falling.

    There is also the reality that house prices are exploding in Canada, Australia, Hong Kong, New Zealand, Great Britain and parts of the USA.

    How to explain such a sustained and international explosion in house prices with dainty economic theories (that do not allow for huge structural impediments)?

    What do such housing costs mean for the large majority of populations in those nations?

    I think all of the mentioned nations could have more housing supply, higher wages and bigger GDPs and better living standards.

    That's quite a fell swoop, no?

    Just get rid of property zoning.


    You may know about CoreLogic. I ran across them in my pen-for-hire business, on mortgage data, of which they have a Niagara. I guess they do house data too.

    1. "Over One-Third of U.S. Top Cities Are Overvalued in Terms of Current Housing Stock"

      O RLY!


    HK to get more pricey?