Robots Don’t Eat Sugar: Productivity Growth and Sugar Consumption Now No Longer Growing Together

In recent years what was once seen as crackpot economics has now become close to conventional wisdom: the notion that productivity costs jobs. Economists call this the lump of labor fallacy. As ITIF has written here, here, and here, it’s clear that the jobs problem of today has nothing to do with productivity and that we should not worry about productivity reducing the number of jobs.

But that has not stopped many talking heads and experts from opining that yes indeed, productivity kills jobs. One graph that has gotten and continues to get widespread attention is from Andrew McAfee and Erik Brynjolfsson’s book, The Second Machine Age, that shows that “productivity and employment have become decoupled.” [i]

But as any first-year statistics course will teach you, correlation does not prove causation. In fact, it is easy to get spurious correlations. Here’s one: The divorce rate in Maine is almost perfectly correlated with the per capita consumption of margarine.

In Brynjolfsson’s case, the relationship being examined merely shows two variables that happen to be increasing from 1970 to 2000, but there is no feasible underlying argument about how one trend should be able to impact the other.

To demonstrate this point, Brynjolfsson’s chart is replicated below, but with daily sugar consumption per capita added. Sugar consumption also appears to be a good fit for labor productivity throughout the 1980s and 1990s, before it falls off in the 2000s. However, it clearly makes no sense to attribute any meaning to this correlation, or suggest that the divergence of these trends around the turn of the millennium has any meaning.

Figure 1: Growth of private employment, labor productivity, and sugar calories consumed daily per capita, 1972-2012, (1997=100) [ii]

So what does impact total employees? Unsurprisingly, the answer is simple demographics—the private employment statistic quoted above grows in step with U.S. population between 25-54 years of age. With retiring baby boomers in the 2000s and the end of the increase in female labor force participation, the rate of labor force growth slowed down dramatically, having absolutely nothing to do with productivity growth.

Figure 2: U.S. productivity, civilian labor force participation and working age population (index, 1981 = 100) [iii]

Employment statistics are impacted by a number of things, but predominately are a function of unemployment rates and population. Labor productivity is actually linked to higher levels of unemployment throughout history. The downturn in employment from 2007 to 2010 is clearly linked to the Great Recession. Blindly asserting that the economy is broken because two indicators that both happened to be growing over a 30-year period is frankly misleading.

—

[i] Andrew McAfee and Erik Brynjolfsson, The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies, 2014, page 165, table 11.1.

[ii] Bureau of Labor Statistics, Employment, Hours, and Earnings from the Current Employment Statistics survey (National), Major Sector Productivity and Costs, Labor Productivity and Private Employment (1972-2012; accessed August 4, 2015); http://data.bls.gov/; United States Department of Agriculture, Food Availability (per capita) Data System (Loss adjusted food availability, calories, sugar, 1972-2012; accessed August 4, 2015); http://www.ers.usda.gov/data-products/food-availability-(per-capita)-data-system.aspx.

[iii] Federal Reserve Bank of St. Louis, FRED Graph (PAYEMS, OPHPBS, LNU00000060; accessed August 29, 2013), http://research.stlouisfed.org/fred2/graph/?g=m4O.