In a new paper, Don’t Blame the Robots: Assessing the Job Polarization Explanation of Growing Wage Inequality, Economic Policy Institute President Lawrence Mishel, Center for Economic and Policy Research senior economist John Schmitt, and EPI economist Heidi Shierholz argue that that technological change cannot explain the wage patterns of the last three decades, especially those of the 2000s.
Some economists believe that technological change in how we produce goods and services has been expanding employment at the top—for lawyers, engineers, and computer programmers, for example—and at the bottom—for janitors, food preparation home health aides, and other low-wage service workers—while actively undermining employment opportunities in the middle. However, Mishel et al demonstrate that occupational employment trends actually provide little insight into the growth of wage inequality or wage deterioration among blue- and white-collar workers.
“Recent news coverage and public policy discussions of the labor market have focused on occupational employment growth patterns and the coming of the robots to explain the growth of wage inequality,” said Mishel. “Our focus needs to be on the widespread deterioration of wages, benefits and job quality in jobs at every skill and education level, rather than tracking the contours of employment shifts across occupations.”
The authors show that technological change has been reducing employment in middle-wage jobs since at least 1950, leading to less employment in blue-collar and administrative occupations and more employment in managerial, technical and professional occupations. However, before the late 1970s, this was associated with rapidly rising wages and flat or declining wage inequality, undermining the case for technology as the driver of the rising wage inequality and the eroding middle class wages that the labor market has experienced since the late 1970s.
The authors note that technology does shape occupational employment patterns over time, but demonstrate that such occupational employment shifts do not materially affect the wage gaps between occupations, and have little impact on overall wage inequality. Thus, the connections between technology-based changes in occupational employment and changes in the wage structure are extremely weak. For example, the employment share in jobs relatively unaffected by automation, such as low-wage service occupations—such as janitors, child care, food prep, health aides —was stable in the 1990s, but wages in these jobs markedly increased. Meanwhile, employment in these same occupations expanded more in the 2000s than in any of the last six decades, but real wages in the occupations actually declined. In fact, the paper finds that the large majority of the increase in wage inequality over the last three decades has occurred within given occupations and not between occupations, especially in the 2000s.
“Our research shows that how quickly—or slowly—employment in an occupation has grown is not strongly related to how quickly—or slowly—wages have grown in that occupation,” said Shierholz. “If we want to know about the conditions of workers in low-wage occupations, we should focus on the factors suppressing their wage growth, and not on their recent occupational employment trends.”
Key findings of the paper include:
- Though much discussed, the phenomenon of job polarization—the expansion of occupational employment at the top and bottom of the wage scale but erosion among middle wage occupations—did not occur in the 2000s before the financial crisis. In particular, occupations paying more than average did not expand their share of employment between 2000 and 2007. Job polarization, therefore, cannot be considered a driver of the growth in wage inequality that occurred in the 2000s.
- Lower-wage service occupations, which are a recent focus of the job polarization story, have remained a small (less than 15 percent of all employment) and relatively stable share of total employment since the 1950s, though they grew in importance in the 2000s.
- Employment in all low-wage occupations, taken together, has been stable for the last three decades, representing a 21.1 percent share of total employment in 1979, 19.7 percent in 1999, and 20.0 percent in 2007. Employment trends in these low-wage occupations have not been a major factor determining the growth or decline in their wages.
- Changes in occupation employment patterns overall have not driven changes in the wage gaps between occupations, and have had even less impact on the overall wage structure. Therefore, even as technology alters occupation employment patterns it does not necessarily drive wage levels of occupations or overall wage inequality.
- Wage trends within occupations are much more important to wage inequality than changes across occupations, so focusing on occupation trends (either wage or employment) provides only limited insights into labor market dynamics.