The Meritocracy Trap

by Daniel Markovits

  workers with JDs and BAs: See Michael Simkovic, “In Law Firms, Lawyers and Paralegals Prosper While Secretarial Jobs Disappear,” Brian Leiter’s Law School Reports, April 1, 2016, accessed November 18, 2018, http://taxprof.typepad.com/taxprof_blog/2016/04/simkovicin-law-firms-lawyers-and-paralegals-prosper-while-secretarial-jobs-disappear.html. As automation moves up the skill hierarchy, the least elaborately trained lawyers will also come increasingly under pressure.

  more intricate and inventive designs: See Richard Sennett, The Corrosion of Character: The Personal Consequences of Work in the New Capitalism (New York: W. W. Norton, 2011), 73. See also Stanley Aronowitz and Willia DiFaxio, The Jobless Future: Sci-Tech and the Dogma of Work (Minneapolis: University of Minnesota Press, 1995), 110.

  within traveling range: See Sherwin Rosen, “The Economics of Superstars,” American Economic Review 71, no. 5 (December 1981): 845–58; and Frank and Cook, The Winner-Take-All Society.

  made nearly $100 million: See “The World’s Highest Paid Celebrities 2017,” Forbes, June 12, 2017, accessed November 18, 2018, www.forbes.com/celebrities/list/#tab:overall.

  more than their midcentury counterparts: Mickey Mantle, for example, made less than $1 million (2015 dollars) per year at midcentury. Michael Haupert, “MLB’s Annual Salary Leaders Since 1874,” Society for American Baseball Research, December 1, 2016.

  more than backup singers: Pam McCallum, “The Average Salary of a Back-Up Singer,” Sapling, June 17, 2011.

  the NBA’s development league: Michael McCann, “The G-League: 12 Takeaways on NBA’s New Deal,” Sports Illustrated, February 14, 2017, accessed November 18, 2018, www.si.com/nba/2017/02/14/nba-gatorade-g-league-deal-adam-silver-takeaways.

  television scriptwriters: “How Much Does a Television Writer Make in the United States?,” Sokanu, accessed October 10, 2018, www.sokanu.com/careers/television-writer/salary/.

  make today: More generally, between 1982 and 2003, the share of total industry revenues captured by the top 1 percent of pop stars more than doubled. See Victor Ginsburgh and David Throsby, eds., Handbook of the Economics of Art and Culture, Volume 1 (Amsterdam: Elsevier, 2006), 684. The ratios of CEO pay to the pay of the next highest-paid employees of a firm, and the ratio of the highest-paid 10 percent of CEOs to the median CEOs, have similarly increased. See Carola Frydman and Raven E. Saks, “Executive Compensation: A New View from a Long-Term Perspective, 1936–2005,” Review of Financial Studies 23, no. 5 (February 2010), http://web.mit.edu/frydman/www/trends_rfs2010.pdf.

  plus substantial benefits: See Reich, Supercapitalism, 89–90. Note that Walmart’s profits on sales are roughly 3.5 percent, or about $6,000 per employee (reported for 2005). This entails that although Walmart might pay its workers appreciably more and still generate profits for its shareholders, it would not afford to pay anything approaching GM’s old wages. Once again, rising economic inequality more significantly reflects deep structural shifts in both production technologies and returns to skill than new exploitation or advantages taking by capital over labor.

  trend toward robot production is accelerating: U.S. investment in robotics more than doubled from 2014 to 2015. See Richard Waters and Kana Inagaki, “Investment Surge Gives US the Early Lead in Rise of the Robots,” Financial Times, May 3, 2016, accessed November 18, 2018, www.ft.com/content/87f44872-1080-11e6-91da-096d89bd2173, reporting study by venture capital research group CB Insights. Hereafter cited as Waters and Inagaki, “Investment Surge.” See also Research Brief, “Robots R’ Us: Funding and Deal Activity to Robotics See New Highs in 2015,” CB Insights, March 23, 2016, accessed November 18, 2018, www.cbinsights.com/research/robotics-startups-funding/. And the worldwide market for production robots alone is growing at a compound annual rate of 17 percent, while annual patent filings related to robotics have tripled in the past decade. Richard Waters and Tim Bradshaw, “Rise of the Robots Is Sparking an Investment Boom,” Financial Times, May 3, 2016, accessed November 18, 2018, www.ft.com/content/5a352264-0e26-11e6-ad80-67655613c2d6. See also Waters and Inagaki, “Investment Surge,” reporting study by economic research group IDC.

  in Europe and Asia: See International Federation of Robotics, World Robotics Report 2016, September 29, 2016, Figure 2.9, https://ifr.org/ifr-press-releases/news/world-robotics-report-2016, and James Carroll, “Industrial Robots in the United States on the Rise,” Vision Systems Design, September 6, 2016, www.vision-systems.com/articles/2016/12/industrial-robots-in-the-united-states-on-the-rise.html.

  since the late 1970s: Total manufacturing employment peaked in the late 1970s, when roughly 19.5 million Americans held manufacturing jobs. See, e.g., Martin Neil Baily and Barry P. Bosworth, “U.S. Manufacturing: Understanding Its Past and Its Potential Future,” Journal of Economic Perspectives 28, no. 1 (Winter 2004): 3–26, 12, Figure 2. Hereafter cited as Baily and Bosworth, “U.S. Manufacturing.”

  Since then, domestic manufacturing employment has fallen steadily. By 1992 the sector sustained only 16.5 million jobs, and by 2012, manufacturing employment had fallen to below 12 million, although it would subsequently recover to just over 12 million. See also Bureau of Labor Statistics, “Employment, Hours, and Earnings from the Current Employment Statistics Survey (National),” extracted June 22, 2017, https://data.bls.gov/timeseries/CES3000000001.

  than it currently provides: In the late 1960s, manufacturing accounted for nearly 25 percent of all U.S. employment. See Baily and Bosworth, “U.S. Manufacturing,” 4, Figure 1. Today, the U.S. civilian labor force comprises roughly 160 million persons, of which roughly 150 million are employed. See Bureau of Labor Statistics, The Employment Situation—May 2017, Summary Table A, https://www.bls.gov/news.release/archives/empsit_06022017.pdf. Twenty-five percent of 150 million is 37.5 million. On the other hand, total manufacturing today employs perhaps 12 million persons. See Baily and Bosworth, “U.S. Manufacturing,” 12, Figure 2.

  increased by 44 percent: This entails that more than 100 percent of the overall decline in manufacturing employment comes from the mid-skilled, non-college-educated segment of the industrial workforce. See Robert Shapiro, “Robotic Technologies Could Aggravate the U.S. Problem of Slow Jobs Growth,” Daily Beast, July 19, 2013, accessed November 18, 2018, www.thedailybeast.com/articles/2013/07/19/robotic-technologies-could-aggravate-the-u-s-problem-of-slow-jobs-growth.html. See also Manufacturing Institute, “Percent of Manufacturing Workforce by Education Level,” April 2014, www.themanufacturinginstitute.org/Research/Facts-About-Manufacturing/Workforce-and-Compensation/Workforce-by-Education/Workforce-by-Education.aspx, and Elka Torpey, “Got Skills? Think Manufacturing,” Bureau of Labor Statistics, June 2014, www.bls.gov/careeroutlook/2014/article/manufacturing.htm.

  The United States is not exceptional in this regard. In each of the seven largest advanced economies—Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States—the share of manufacturing jobs held by white- rather than blue-collar workers increased in both the 1980s and the 1990s. See Mariacristina Piva, Enrico Santarelli, and Marco Vivarelli, “The Skill Bias Effect of Technological and Organizational Change: Evidence and Policy Implications,” Research Policy 34 (2005): 141–57, 143.

  even as its share of employment declined: Baily and Bosworth, “U.S. Manufacturing,” 4, Figure 1. See also YiLi Chien and Paul Morris, “Is U.S. Manufacturing Really Declining?,” Federal Reserve Bank of St. Louis On the Economy Blog, April 11, 2017, accessed January 28, 2019, www.stlouisfed.org/on-the-economy/2017/april/us-manufacturing-really-declining; Peter Wehner and Robert Beschel Jr., “How to Think About Inequality,” National Affairs 11 (Spring 2012), accessed November 18, 2018, www.nationalaffairs.com/publications/detail/how-to-think-about-inequality; Rex Nutting, “Think Nothing Is Made in America? Output Has Doubled in Three Decades,” MarketWatch, March 28, 2016, www.marketwatch.com/story/us-manufacturing-dead-output-has-doubled-in-three-decades-2016-03-28; Harding, “Technology Shakes Up US Economy.”

  Man
ufacturing has sustained its output share almost entirely on account of massive increases in productivity and output in the computer and electronics industry, which hires disproportionately high-skilled workers, increasingly to design and implement robotic production. See Baily and Bosworth, “U.S. Manufacturing,” 3–26. For example, a computer’s most valuable component, the motherboard, is today largely robot-made. See Catherine Rampell, “When Cheap Foreign Labor Gets Less Cheap,” New York Times, December 7, 2012, accessed November 18, 2018, https://economix.blogs.nytimes.com/2012/12/07/when-cheap-foreign-labor-gets-less-cheap/?partner=rss&emc=rss&_r=0.

  increased by over 15 percent: See Harding, “Technology Shakes Up US Economy.”

  it was sold to Facebook for $1 billion: On Kodak, see Susan Christopherson and Jennifer Clark, Remaking Regional Economies: Power, Labor, and Firm Strategies in the Knowledge Economy (New York: Routledge, 2007), 57–84.

  On Instagram, see Scott Timberg, “Jaron Lanier: The Internet Destroyed the Middle Class,” Salon, May 12, 2013, accessed November 18, 2018, www.salon.com/2013/05/12/jaron_lanier_the_internet_destroyed_the_middle_class/. Hereafter cited as Timberg, “Internet Destroyed.”

  Instagram’s tiny super-skilled workforce can meet the firm’s staffing needs only because it works alongside the mass of people who use its technology to stage, capture, process, and print images and thus constitute, in structure, a low-skilled labor force, even as they conceive of themselves not as producers but rather consumers. Jaron Lanier thus observes that lots of people—the site’s users who contribute content and even formatting—“work” at Instagram, in a sense. See Timberg, “Internet Destroyed.”

  This pattern, incidentally, is far from unique. It has occurred before, including in connection with image production: photography, as Lanier also notes, itself enabled a mass of low-skilled snapshot takers to displace mid-skilled painters and illustrators. See Timberg, “Internet Destroyed.” The same pattern arises in other sectors also: to name but one example, Ikea’s customers also function, structurally speaking, as a mass of unskilled furniture assembly labor.

  fabulously wealthy, of course: Facebook itself employs far fewer people than Kodak ever did, and several of them have become far richer than Kodak’s founder, George Eastman, ever was.

  mostly since 2000: The share of nonroutine cognitive jobs in the labor market overall grew by 8.2 percent between 1982 and 1992, 10.6 percent between 1992 and 2002, and 14.9 percent between 2002 and 2017. The share of nonroutine manual jobs grew by 0.8, 1.3, and 9.4 percent in each of these decades. And the share of routine jobs fell by 4.5, 6.7, and 13.7 percent. These data come from Nir Jaimovich and Henry E. Siu, “Job Polarization and Jobless Recoveries,” 8, Figure 3, forthcoming in Review of Economics and Statistics, accessed November 18, 2018, http://faculty.arts.ubc.ca/hsiu/pubs/polar20180903.pdf. Hereafter cited as Jaimovich and Siu, “Job Polarization and Jobless Recoveries.”

  increased by more than a third: Nonroutine cognitive jobs have increased from 29 to 40 percent of all jobs, routine jobs have fallen from 56 to 42 percent, and nonroutine manual jobs have increased from 15 to 18 percent. These conclusions simply cumulate the decadal shifts reported in the prior note.

  is today nearly 20 percent: Cappelli, The New Deal at Work, 159–60. Cappelli adds that on the midcentury model, the high/low-status fault line in the workplace lay between exempt and nonexempt workers under the Fair Labor Standards Act. See Cappelli, The New Deal at Work, 236–37. Today, the fault line lies between top managers and everyone else.

  falling middle-class incomes: See Jeffrey P. Thompson and Elias Leight, “Do Rising Top Income Shares Affect the Incomes or Earnings of Low and Middle-Income Families,” B.E. Journal of Economic Analysis & Policy 12, no. 1 (2012): 26. Thompson and Leight deploy cross-state comparisons, which reveal that states with more rapidly rising top income shares experience greater declines in middle incomes. The preferred specification associates a 3.5 percent increase in the top 1 percent’s income share between 1979 and 2005, with a statistically significant 3.5 percent reduction over the same period in incomes of families between the 35th and 70th percentiles of the overall distribution. (The study also interestingly suggests, in line with the by now familiar pattern, that rising top incomes have only a smaller and less statistically significant negative impact on incomes at the bottom of the distribution, in families between the 5th and 30th percentiles.)

  throughout the world’s rich societies: One study found significant hollowing out of mid-skilled jobs in every European country but Portugal. And another study, compiled by the OECD using data aggregated across twenty-four countries, found that from 1998 to 2009, occupations in the highest and lowest quartiles by average skill scores experienced employment growth (of nearly 25 and 2 percent respectively), while occupations with the two middle skill quartiles experienced employment declines (of 1 percent for the second-highest skill quartile and 15 percent for the second-lowest skill quartile). Indeed, according to the OECD, “In half the OECD countries for which data are available, the loss of jobs associated with a medium level of education was greater than the loss of jobs associated with a low level of education.” See Marten Goos et al., “Explaining Job Polarization in Europe: The Roles of Technology, Globalization and Institutions,” Centre for Economic Performance Discussion Paper No. 1026 (November 2010); Daniel Oesch and Jorge Rodriguez Menes, “Upgrading or Polarization? Occupational Change in Britain, Germany, Spain, and Switzerland, 1990–2008,” MPRA Paper No. 21040 (January 2010); but see Enrique Fernandez-Macias, “Job Polarization in Europe? Changes in the Employment Structure and Job Quality, 1995–2007,” Work and Occupations 39, no. 2 (2012); Rachel E. Dwyer and Erik Olin Wright, “Job Growth and Job Polarization in the United States and Europe, 1995–2007,” in Transformation of the Employment Structures in the EU and USA, ed. Enrique Fernandez-Macias, Donald Storrie, and John Hurley (Basingstoke: Palgrave Macmillan, 2012), 49, Appendix B, Table B1.6; Alexandra Spitz-Oener, “Technical Change, Job Tasks, and Rising Educational Demands: Looking Outside the Wage Structure,” Journal of Labor Economics 24, no. 2 (April 2006): 235–70; Goos and Manning, “Lousy and Lovely Jobs.”

  even as mid-wage employment declined: See NELP, National Employment Law Project, The Low-Wage Recovery and Growing Inequality (New York: NELP, 2012).

  greater shares of gains than of losses: Low-wage jobs accounted for 21 percent of job losses in the recession and 58 percent of job growth in the recovery, high-wage jobs accounted for 19 percent of losses and 20 percent of growth, and mid-wage jobs accounted for fully 60 percent of losses but just 22 percent of growth. See National Employment Law Project, The Low-Wage Recovery, 1. See also Robert Reich, “The Hidden Price of Your Amazon Shopping Spree: Skyrocketing Unemployment,” Salon, December 4, 2013, accessed November 18, 2018, www.salon.com/2013/12/04/robert_reich_healthcare_gov_is_not_the_website_we_should_be_worrying_about_partner/.

  will all be either low- or super-skilled: The fastest-shrinking job categories include mid-skill work such as apparel, leather, and allied manufacturing; tobacco manufacturing; the Postal Service; other federal employment; and manufacturing and reproducing magnetic and optical media. The fastest-growing job categories include low-skill work such as home health care; outpatient health care; health office work; facilities support; and ambulatory care. They also included high-skill work such as management, scientific, and technical consulting; software publishing; computer systems design; and securities, commodity contracts, and financial investments. The only fast-growing mid-skill job categories are nursing and medical and diagnostic lab work. See also Max Nisen, “Ten American Industries That Are Going to Boom in the Next Decade,” Slate, December 28, 2013, accessed November 18, 2018, https://slate.com/business/2013/12/booming-industries-for-the-next-decade.html.

  Academic observers similarly predict that nearly half of jobs might be made redundant within twenty years. The jobs most likely to be displaced are routine or routinizable and ther
efore mid-skilled: loan officers, receptionists, paralegals, retail salespersons, and taxi drivers. The jobs least likely to be displaced are all fluid and require social perception and creative intelligence: reporters, physicians, lawyers, teachers, and doctors. Carl Benedikt Frey and Michael A. Osborne, “Job Automation May Threaten Half of U.S. Workforce,” Bloomberg, March 12, 2014, accessed November 18, 2018, www.bloomberg.com/graphics/infographics/job-automation-threatens-workforce.html.

  displaced by automation by 2030: James Manyika et al., “Jobs Lost, Jobs Gained: What the Future of Work Will Mean for Jobs, Skills, and Wages,” McKinsey Global Institute, November 2017, accessed October 26 2018, www.mckinsey.com/featured-insights/future-of-work/jobs-lost-jobs-gained-what-the-future-of-work-will-mean-for-jobs-skills-and-wages.

  even a sea change: Other explanations for labor market polarization that emphasize forces besides technological innovation of course also exist. Candidates include globalization, declining union membership, and changes in tax policy. All these causes certainly contribute to the demise of the middle class. But none has been as consequential as automation, some (most notably globalization) are themselves caused by automation, and all will have much small effects than automation on middle-class employment going forward.

  isolates super-skilled workers from all others: On the midcentury model, the high/low-status fault line in the workplace lay between exempt and nonexempt workers under the Fair Labor Standards Act. Today, the fault line lies between top managers and everyone else. Cappelli, The New Deal at Work, 236–37.

  provided in universities: Workplace training makes sense where it can be delivered in small and regular increments, throughout a worker’s career. The midcentury labor market supported this model, by offering a many-tiered hierarchy of jobs, with each rung of the professional ladder in close proximity to the ones below and above it. University-taught and degree-based professional education, by contrast, must be delivered in large doses, and only irregularly, as workers interrupt their jobs to receive extended spells of full-time education. Professional school makes sense where labor markets offer job hierarchies with fewer tiers, separated by wider gaps. The shift from workplace-centered to university-based professional training thus suggests that there has been a parallel and profound change in the labor market: a many-tiered hierarchy of jobs has had its middle rungs removed, leaving jobs at the bottom and jobs at the top, separated by a wide chasm.