In our September 21st post, “Whatever It Takes,” we examined today’s low growth in productivity,[1] which has accounted for about 50% of the substandard growth rate in GDP since the Great Recession of a decade ago.[2] During his September 2018 speech at the annual Jackson Hole Symposium, Fed Chairman Jerome Powell offered some optimism in his hope for a “Fourth Industrial Revolution.” This is a common refrain that assumes the next several decades will see advances in 3-D printing, autonomous vehicles, robots, artificial intelligence, and robotics that will rekindle productivity growth.

Among his fellow optimists are Erik Brynjolfsson and Andrew McAfee, whose best-selling book The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies (2014) is both intriguing and balanced. The authors conclude that the economic efficacy of the innovations above will be accompanied by massive job destruction and related social consequences. Interestingly, Joel Mokyr—friend, fellow professor, and intellectual sparring partner of Robert Gordon at Northwestern, a productivity skeptic—declares unabashedly that “based on rapidly improving scientific insights, technological breakthroughs have the potential to change life in the foreseeable future as much as they did in the century and a half after the Civil War.” Mokyr’s essay “The Past and Future of Innovation: Some Lessons from Economic History” appears in the journal Explorations in Economic History (2018).[3]

At first blush, Gordon appears to be the proverbial skunk at the garden party. Where Gordon differs with Mokyr et al. is not on the pace of innovation, which he thinks will be impressive, but rather on the extent to which it will contribute to productivity. Moreover, he argues that the magnitude of effect that the likes of electricity and the internal-combustion engine had on human existence—let alone their enormous contribution to productivity growth—were life-changing and have no comparable analog in the much-anticipated Fourth Industrial Revolution. Gordon’s 2016 book, The Rise and Fall of American Growth: The U.S. Standard of Living since the Civil War, is a tour de force through the greatest industrial age in the history of the planet.

Zeroing in on the four categories of innovation that offer the most immediate prospect for productivity enhancement over the next 25 years—3-D printing, autonomous (driverless) vehicles, robots, and artificial intelligence (AI)—Gordon argues that progress thus far suggests that the impact on productivity growth and job destruction will be “gradual and evolutionary, not sudden and revolutionary.” 3-D printing, also known as additive manufacturing, at this stage is better suited for small-batch processing than manufacturing 17 million cars and light trucks a year. Though gains will be measured in figurative inches and not miles, autonomous vehicles are clearly in our future. The first job classification likely to be displaced will be the long-haul trucker. Despite the economic appeal, the FAA has not even begun to consider allowing commercial airline flights with a single pilot.

Robots, which were first introduced to manufacturing in 1961, have made few inroads outside of manufacturing and warehousing. They are simply not adept at many of the routine tasks performed by humans. Artificial intelligence, where machines mimic human learning and problem solving, was founded as an academic discipline in 1956. Its history has seen waves of optimism followed by disappointment and loss of funding, all the while surrounded by ethical controversy similar to that facing the pioneers in genetic engineering. I am no technophobe. I benefit personally from these advances as AI allows software on my PC to understand my speech.[4] Still, while progress has been steady, it has been incremental.

The future is conjecture, the past is fact. The productivity of labor—the percentage change in output per hour—has been in decline: from an annual rate of 2.82% (1920-1970) to 1.75% (1970-2006) to 0.97% (2006-2016). Given the long-term downtrend in productivity growth, is it not fair to ask to whether the malady is structural?

The Nemesis of Productivity Growth

One structural factor to declining productivity is the slowing rate of increase in educational attainment.

This is of overarching social and economic importance. To be clear, we are discussing the rate of growth in education. Since there’s a maximum of persons to be educated, secular growth rates cannot forever increase. The problem is insidious and so glacial as to be beyond the public consciousness. It rarely captures the headlines. Solutions to spur growth in education—like student loans—have arguably been counterproductive.

To set the scene, the trajectory of educational attainment reached an inflection point around 2000. The annual growth rate of the percentage of the population to complete high school was 3.3% per year from 1915 to 1980 but only 0.2% from then until 2016. While less dramatic, the four-year college degree completion growth rate declined from 3.7% to 1.3% over the same periods. Though beyond the scope of this post, the high correlation between educational attainment and productivity growth can actually be calculated.[5]

Zooming in on the percentage of young people seeking four-year college degrees, for reasons of both demand and supply, the decline in the rate of increase is likely to continue. Despite cyclical fluctuations in education inversely correlated with demand for workers, Beaudry et al.[6] puts forth a plausible thesis for structural decline. The buildout and diffusion of digital infrastructure driven by the short-lived “Third Industrial Revolution,” from roughly 1996 to 2006, has been subject to the boom-and-bust cycle common to most investment models.

Further, in addition to the well-publicized structural decline of jobs in routine occupations and manufacturing, Beaudry posits that after two decades of growth in the demand for occupations high in cognitive skills—long a key driver of the U.S. labor market—the U.S. economy reversed direction around 2000 and experienced a decline in the demand for such skills due to the maturation of that infrastructure. The law of comparative advantage explains why a reduction in the demand for more highly educated workers caused the employment rate in the U.S. to contract. Higher skilled workers, displaced from cognitive occupations, moved down the occupational ladder and pushed many lower skilled workers out of the labor market altogether. This process has been going on since 2000, but the housing bubble and construction boom between 2003 and 2006 masked some of the effects that only became fully apparent after the financial crisis.

Particularly in the STEM fields (science, technology, engineering, math), students have been under-enrolled, constricting the supply of workers that businesses still do need. The credentials for such positions are so specific, however, that the rising cost of higher education and the increasing burden of student debt, not to mention the state of secondary education, contribute to the conundrum. From 1975 to 2015 the price of higher education increased at a 7.1% rate compared with 3.4% for the CPI. Student-loan debt has quintupled since 2000 to $1.5 trillion.[7] As a result, the supply/demand curve shifted downward. After rising rapidly from 1980 to 2000, the wage premium for college-educated workers grew by only a small amount between 2000 and 2010 and then was flat during the economic recovery from 2010 to 2015, the latest data point. This is not true for every occupation, but for the average high school graduate economist Tim Harford deems it rational that a rising number of individuals consider college a dicey financial investment.[8]

The Frown Behind the Smiley Face

The most recent labor report, trumpeting the lowest unemployment rate in 50 years, diverted attention away from a more troubling statistic. The red curve on the FRED[9] chart below, the civilian unemployment rate, was featured in our last post. Here, we added another curve, the Civilian Employment-Population Ratio, which purports to measures the proportion of the U.S. working-age[10] population (16 and older) that is employed. The ratio is used to evaluate the ability of the economy to create jobs and therefore is used in conjunction with the unemployment rate for a general evaluation of the quality[11] of the labor market. Having a high ratio means that a greater proportion of the population of working age is employed, which in general will have a positive effect on GDP per capita. As per the title of this post, without robust and sustainable rates of growth in productivity, its effects—real growth in GDP, wages, profits, and investment—are unlikely to remain elevated.

employment curves

When you view the chart longitudinally, you will notice that from the mid-1960s[12] to 2000, the Civilian Employment-Population Ratio maintained an upward trend, interrupted only by recessions. Consistent with what was written above, although not limited to those arguments, the ratio did not repeat that trend following the recession of 2001 or the Great Recession at the end of the decade. Keep in mind that the oldest Baby Boomers reached age 65 in 2011.

If changes to growth in productivity are structural, low growth of GDP would follow. In that environment, the case for increased investment spending is weak. Moreover, if savings are the source of investment, neither the government, the consumer, nor even corporations are prepared to take that step regardless of GDP growth. Today’s risk assets are not priced for such scenarios.

 

[1] Labor productivity refers to the total economy, i.e., real GDP/total hours of work.

[2] The correlation between growth in GDP and growth in productivity is high: Recall from our post, “Whatever It Takes,” that U.S. economic growth slowed by more than half from an average of 3.2% per year during 1970–2006 to only 1.6% during 2006–16. Productivity growth followed a similar pattern, the annual rate of growth declining from 1.75% to 0.35%, respectively.

[3] Joel Mokyr, “The Past and Future of Innovation: Some Lessons from Economic History,” in Explorations in Economic History, Vol. 69 (2018), pages 13–26.

[4] The original Dragon Dictate (1990) was costly at $9,000 compared to an order of magnitude better version today costing $149.99.

[5] Robert J. Gordon, “Why Growth Is So Slow When Innovation Is Accelerating?” NBER Working Paper No. 24552 (April 2018), 13.

[6] Paul Beaudry, David Green, and Ben Sand, “The Great Reversal in the Demand for Skills and Cognitive Tasks,” Journal of Labor Economics 34 (June 2016), S199–S247.

[7] A federal program antithetical to its intended goals? Moreover, graduates saddled with debt are aggravating the long-term decline in the share of newly formed businesses.

[8] Tim Harford, The Logic of Life: The Rational Economics of an Irrational World, Random House Publishing Group, 2008. Kindle Edition, Location 121.

[9] Economic charts provided by the Federal Reserve Bank of St. Louis. Available to everyone.

[10] Literature is inconclusive as to what constitutes the upper limits of working age.

[11] From labor’s perspective, the Civilian Employment-Population Ratio does not indicate whether workers have jobs in alignment with their skills, the number of hours (or jobs) worked per person, and whether wages or salaries fit within the admittedly imprecise definition of a “quality” labor market.

[12] After the birth-control pill gained acceptance in the mid-1960s, the labor force participation rate of women rose from 37% to a peak of 60% in 2000.

Leave a Reply