Autor 2010 — The Polarization of Job Opportunities in the U.S. Labor Market

Note: This is a policy report jointly released by the Center for American Progress and The Hamilton Project (Brookings Institution), April 2010. It is not a peer-reviewed journal article; it synthesizes Autor's prior empirical work (especially Autor, Levy, and Murnane (ALM) 2003, Autor, Katz, and Kearney 2006/2008) for a policy audience.

Summary

Autor (2010) documents the "polarization" of U.S. job opportunities since 1979: employment and earnings have grown at both the high-skill and low-skill ends of the occupational distribution while collapsing in the middle. Middle-skill jobs — clerical, administrative, production, and operative occupations — fell from $57.3\%$ of employment in 1979 to $48.6\%$ in 2007 and $45.7\%$ by 2009. The same pattern is documented in all 16 European Union (EU) countries 1993–2006. The report argues that routine-biased technological change (RBTC) — computerization substituting for rule-based tasks while complementing abstract reasoning and leaving interpersonal manual tasks intact — is the primary mechanism. The Great Recession accelerated polarization ($r = 0.76$ between pre-recession and within-recession occupational employment growth), and real wages for non-college workers have fallen sharply (high school (HS) male $-12\%$, HS dropout $-16\%$ real, 1979–2007).

Key Claims

• Middle-skill hollowing: Employment share of middle-skill occupations: $57.3\%$ (1979) $\to$ $48.6\%$ (2007) $\to$ $45.7\%$ (2009). High-skill professional/technical/managerial and low-skill service occupations gained.
• RBTC mechanism: The task-based framework (Autor, Levy, Murnane 2003) — computers substitute for routine cognitive (clerical, data entry) and routine manual (assembly, production) tasks; complement non-routine abstract (analysis, management) tasks; cannot replace non-routine manual (service, personal care) tasks. Middle-skill jobs are intensive in routine tasks.
• International breadth: All 16 EU countries in the EU-15 + Norway sample show employment polarization 1993–2006. The pattern is not U.S.-specific.
• Polarizing recession: Within-recession occupational employment growth (2007–2009) correlates $r = 0.76$ with pre-recession growth (2001–2007) — recessions accelerate ongoing polarization rather than disrupting it.
• Gender asymmetry: Female workers shifted upward in the skill distribution across the full period. Less-educated male workers shifted primarily downward into low-skill service work. Men without college degrees have experienced the most severe relative earnings decline.
• College wage premium: College/high school hourly wage ratio: $\approx 1.50$ (1963) $\to$ $1.95$ (2009). The entire rise occurred post-1980; it reflects both rising college wages and declining non-college wages simultaneously.
• Supply slowdown: White male 4-year college attainment barely moved ($20\% \to 26\%$, 1970–2008) relative to female gains ($12\% \to 34\%$). The post-Vietnam War cohort effect and low returns in the 1970s explain the male stagnation; slowing supply amplified the college wage premium even as demand kept rising.
• Employment-to-population elasticity: Across demographic groups, a $10\%$ wage decline is associated with a $5.8$ percentage points (pp) decline in employment-to-population ratio (ordinary least squares (OLS) across demographic cells).
• Policy implication: Standard "job training" prescriptions are insufficient; the problem is a structural shift in the demand for skills driven by technology and trade, not temporary skill gaps.

Concepts Introduced or Extended

• Occupational Polarization
• Routine-Biased Technological Change

Entities Mentioned

• David Autor

Quotes

"The U.S. labor market has been undergoing a profound structural change. The structure of job opportunities in the United States has sharply polarized over the past two decades, with expanding job opportunities in both high-skill, high-wage occupations and low-skill, low-wage occupations, coupled with contracting opportunities in middle-skill, middle-wage white collar and blue collar jobs."

"The growth of high and low wage employment at the expense of middle wage employment is not a uniquely American phenomenon. All 16 countries of the EU 15 plus Norway saw employment polarize between 1993 and 2006."

"The impact of the Great Recession on occupational structure has been highly non-neutral. The industries and occupations that contracted most in the recession were those that were already contracting before the recession."

My Take

This is a synthesis report, not original empirical research — the underlying empirical work is in ALM 2003, Autor, Katz, and Kearney (2006/2008), and Autor and Dorn (2013). Its value is as a clear, data-rich articulation of the polarization framework for a policy audience, and as a temporal snapshot of where Autor's thinking stood in 2010, two years before the Autor, Dorn, and Hanson (ADH) China Syndrome papers. The employment-to-population elasticity ($5.8$ pp per $10\%$ wage decline) is a useful summary statistic linking earnings and participation trends. The "polarizing recession" finding — $r = 0.76$ between pre- and within-recession occupational growth — is an important empirical observation showing that the Great Recession was not a random demand shock but an amplification of underlying structural trends. The limitation is that the causal identification is loose: the RBTC mechanism is supported by correlational task-content analysis (ALM 2003), not by an experiment. The China trade shock literature (ADH 2013+) later provided sharper causal identification for the trade component of the same phenomenon. See China Trade Shock.