Earnings Inequality: Stylized Facts, Underlying Causes, and Policy Barry Hirsch Department of Economics Andrew Young School of Policy Sciences Georgia State University Prepared for Atlanta Economics Club luncheon address, Federal Reserve Bank of Atlanta, December 17, 2014
Earnings Inequality: Stylized Facts, Underlying Causes, and Policy Overview: 1. Types of inequality how and why they matter 2. Descriptive evidence on US earnings inequality 3. Underlying causes of rising earnings inequality demand, supply, and institutional forces [Demand] Skill biased technological change (SBTC) [Supply] Slow growth in educated workers (skills) relative to demand [D & S] Globalization: Flows of Goods (trade), Capital (investment, offshoring), People (immigration), and Knowledge [Institutional] Low minimum wages (MW) [Institutional] Decline in private sector unionism 4. Policy implications 1
Focus on wage and earnings differences in the labor market. Economic theory of wage differentials (Adam Smith 1776) In competitive (and non-competitive) labor markets, wage differences (i.e., inequality) arise due to worker and job differences and the interaction of labor supply and demand. Institutions and laws (unions, minimum wages, etc.) also matter. Inequality inevitable and desirable, but large inequalities undermine societies to the extent that they: result from unequal opportunities, are perceived as not fully deserved, and distort political outcomes. Fairness matters. Distinction between (in)equality of opportunity vs. (in)equality of results. 2
What types of inequality are relevant? We will focus on wage or earnings inequality these are tied to labor market outcomes Depending on the question/issue being addressed, we also care about: Household income inequality Wealth inequality (asset wealth vs. human capital wealth) Consumption inequality Other issues: Cross-sectional vs Lifetime inequality (inequality age-related) (total vs. residual inequality) Earnings and income Mobility within and across generations Measurement Multiple measures: Gini coefficient, log variance, percentile ratios 90/10, 90/50, 50/10 No single measure can summarize an entire earnings (or income) distribution Inequality can increase due to more persons at bottom, more at top, fewer in middle 3
Descriptive data on US earnings inequality Primarily from Daron Acemoglu and David Autor, What Does Human Capital Do? A Review of Goldin and Katz s The Race Between Education and Technology, Journal of Economic Literature, 2012. 4
Figure 5a: Trends in Full- Time, Full- Year Weekly Wages Cumulative Log Change in Real Weekly Earnings at the 90th, 50th and 10th Wage Percentiles 1963-2008: Full-Time Full-Year Males 0.2.4.6 1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 10th Percentile 90th Percentile 50th Percentile
Figure 6 Real, Composition-Adjusted Log Weekly Wages for Full-Time Full-Year Workers 1963-2008 Males 0.2.4.6 Composition-Adjusted Real Log Weekly Wages 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 Year HSD SMC GTC HSG CLG Source: March CPS data for earnings years 1963-2008. See note to Figure 1. The real log weekly wage for each education group is the weighted average of the relevant composition adjusted cells using a fixed set of weights equal to the average employment share of each group.
Figure 8 Smoothed Changes in Employment by Occupational Skill Percentile 1979-2007 100 x Change in Employment Share -.05 0.05.1.15.2 0 20 40 60 80 100 Skill Percentile (Ranked by Occupational Mean Wage) 1979-1989 1990-2007 Source: Census IPUMS 5 percent samples for years 1980, 1990, and 2000, and Census American Community Survey for 2008. All occupation and earnings measures in these samples refer to prior year s employment. The figure plots log changes in employment shares by 1980 occupational skill percentile rank using a locally weighted smoothing regression (bandwidth 0.8 with 100 observations), where skill percentiles are measured as the employment- weighted percentile rank of an occupation s mean log wage in the Census IPUMS 1980 5 percent extract. Mean education in each occupation is calculated using workers hours of annual labor supply times the Census sampling weights. Consistent occupation codes for Census years 1980, 1990, and 2000, and 2008 are from Autor and Dorn (2009a).
Figure 9 Percent Change in Employment by Occupation, 1979-2010 -.2 0.2.4.6 Personal Care Protective Service Food/Cleaning Service Production Operators/Laborers Office/Admin Sales Technicians Professionals Managers 1979-1989 1989-1999 1999-2007 2007-2010 Source: May/ORG CPS files for earnings years 1979-2010. The data include all persons ages 16-64 who reported having worked last year, excluding those employed by the military and in agricultural occupations. Occupations are first converted from their respective scheme into 326 occupation groups consistent over the given time period. All non- military, non- agriculture occupations are assigned to one of ten broad occupations presented in the figure.
Figure 10 Change in Employment Shares by Occupation 1993-2006 in 16 European Countries Occupations Grouped by Wage Tercile: Low, Middle, High Percent Change -.15 -.1 -.05 0.05.1.15.2 USA EU Average Italy Austria France Luxembourg Denmark Belgium Spain Germany Sweden UK Greece Netherlands Norway Finland Ireland Portugal Lowest Paying 3rd Highest Paying 3rd Middle Paying 3rd Source: Data on EU employment are from from Goos, Manning and Salomons, 2009a. US data are from the May/ORG CPS files for earnings years 1993-2006. The data include all persons ages 16-64 who reported having worked last year, excluding those employed by the military and in agricultural occupations. Occupations are first converted from their respective scheme into 326 occupation groups consistent over the given time period. These occupations are then grouped into three broad categories by wage level.
Note: Analysis using CPS data and 90/10 percentiles misses wage growth at the very top (public files have top-coded earnings to protect confidentiality). Recent growth in inequality has come disproportionately from top 1% (Piketty and Saez). Thomas Piketty and Emmanual Saez, "Income Inequality in the United States, 1913-1998," Quarterly Journal of Economics, 2003 Wojciech Kopczuk, Emmanual Saez, and Jae Song, "Earnings Inequality and Mobility in the United States: Evidence from Social Security Data since 1937," Quarterly Journal of Economics, 2009 Much of the earnings increase at the very top has been associated with incentive compensation (bonuses, stock options, etc.). Thomas Lemieux, W. Bentley MacLeod, and Daniel Parent, Performance Pay and Wage Inequality, Quarterly Journal of Economics, 2009 5
Why has earnings inequality increased? Demand, supply, and institutional forces [Demand] [Supply] Skill biased technological change (SBTC), increased demand for skills Slow growth in educated workers relative to demand (losing the race between technology and education) [Demand and Supply]: Globalization Flows of goods (trade), capital (investment/ plants), and people (immigration) [Institutional] Low minimum wages (MW) [Institutional] Decline in private sector unionism 6
Simple skill biased technological change (Simple SBTC) and The Race between education (supply) and technology (demand) Think of information technology/computers and other technologies Technology substitutes (decreases demand) for lower skill workers Technology complements (increases demand and productivity) for higher skill workers Inequality is a race between SBTC demand changes and supply of educated workers (Lawrence Katz and Claudia Goldin, The Race between Education and Technology, 2008) This is an over-simplification, but provides a rough approximation of why earnings inequality has increased. Does a good job explaining decreasing and increasing returns to college, decreasing in the 1970s and increasing in the 1980s. Does not explain well what has occurred since the 1990s ( hollowing of the middle). 7
Job task or nuanced SBTC (David Autor, others) from Information Technology (IT) IT is labor saving (decreases labor demand) for routinizable or programmable tasks Production workers in plants (robotics) Information based workers: bank tellers, reservation agents, bookkeepers, etc. IT complements (increases productivity) for non-routinizable abstract or analytical tasks Examples: lawyers, accountants, administrative assistants, architects, economists Note: IT and the Internet may allow analytic tasks to be provided from a distance through outsourcing or telecommuting; e.g., call centers, business accounting IT has little effect on manual, non-programmable tasks delivered in person Examples: hair stylists, child-care workers, landscaping & groundskeepers, physical therapists 8
Autor, Levy, Murnane (2003) Evidence: consistent with Job task SBTC approach Industries in which jobs had high levels of routine tasks in 1960s had substantially higher levels of IT (computer adoption) through 1997 Employment growth through 1998 high in nonroutine jobs and falling in routinizable jobs since 1980 Computers task shifts within jobs college going and patterns of employment Computers explain much of task changes which in turn affect educational attainment through changes in occupational employment and tasks performed within occupations. Principal results Relative decrease in employment & earnings for jobs with routinizable tasks Relative increase in employment & earnings for jobs with non-routinizable tasks Little change in relative earnings for jobs with manual, non-programmable tasks Loss of middle-class jobs. A hollowing out of the middle of the earnings distribution. 9
So let s return to other possible explanations (suspects) for rising inequality Globalization: Movement of goods (trade), capital (investment/plants), people (immigration), knowledge Wage differences have narrowed across countries International trade increasingly important, particularly Chinese trade since 2001 and its effects on manufacturing industries. Also important are increased mobility of capital and off-shoring of production. Note: Inequality within almost all countries has increased over time, but worldwide inequality in incomes across all persons/households has decreased quite a bit. Explanation: Relative earnings and incomes in many developing countries (China and India) have sharply increased lowering worldwide poverty and inequality. Yet within countries more skilled workers have fared well relatively to less skilled. 10
Immigration Immigration in U.S. has increased steadily until Great Recession. 15% of US wage and salary employment is foreign-born. Concentrated in the tails of the skill distribution Many college and graduate degree immigrants who are educated in U.S. and stay Concentration of young, low-skill immigrants, many from Mexico/Central America Theory suggests low-skill immigration should lower wages for low skill workers and raise or have little effect on real wage of higher skill workers. Evidence Negative effect of immigration on low-skill wages, but surprisingly small Increase in overall employment, but little effect on native employment New immigration lower wages of earlier immigrants who most directly compete Immigration the principal cause of rising inequality timing not right Immigration cannot explain sharp deterioration in left tail during the 1980s Large immigration increases in 1990s and 2000s, but left tail held up well. Middleclass jobs deteriorated by were little affected by immigration Immigration flows highly sensitive to job opportunities in US & in source countries. Bottom line: Immigration plays small or modest role in increasing inequality 11
Minimum Wages MW fell during 1980s and has remained low by historical standards MW affects inequality through changes in the lower tail of the distribution Helps explain some of the sharply rising inequality in 1980s, but little since then MW much more important for wage inequality for women than for men Higher minimum wages does little to prevent middle class job and wage erosion Roughly 4% of workers at or below current $7.25 Federal MW 12
Federal Minimum Wage in Nominal (blue) and Real (red) $2014 dollars. Many cities & states depart from $7.25 federal MW. In 2014, San Francisco has highest MW at $10.55. Highest state MW is Washington's at $9.32; Oregon's second at $9.10. 13
Decline in Private Sector Unionism Private sector union density currently below 7% (see figure). Public union density much higher. Roughly half of all union members now work for federal, state, local government In 1940s-1970s we had union governance workplace norm in the industrial sector (manufacturing, construction, and transportation, communications, and utilities). Union governance highly formalized (contractual) with reduced management discretion Unionization decreases wage dispersion/inequality through: Compressing wages top to bottom Standardizing wages (less individualized wage dispersion) through contractual terms tying wages to designated job positions and seniority Limits executive compensation Bottom line: Decline in union density appears to account for roughly 15-20% of increase in male wage inequality in US. Less important for women. 14
Union Membership (Millions) 25 Union Membership among U.S. Wage and Salary Workers, 1973-2013 20 15 10 5 0 1973 1977 1981 1985 1989 1993 1997 2001 2005 2009 2013 Private Public Overall 15
Percent Union Membership 45 40 35 30 25 20 15 10 5 Union Membership Density among U.S. Wage and Salary Workers, 1973-2013 0 1973 1977 1981 1985 1989 1993 1997 2001 2005 2009 2013 Private Public Overall 16
Policy implications difficult to decrease inequality through desirable policies Discourage technological change? No. Changes in technology provide the principal engine for economic and income growth. Increase supply of educated workers? Yes, but difficult to do. We heavily subsidize college and graduate education, yet schooling growth is slow. Large numbers of high school grads start college but complete less than a year. Greater gains might come from investing in pre-school children and improving the quality of primary and secondary schools. Enact trade and investment (capital flow) barriers. Might decrease inequality to some small degree but would weaken growth and real incomes (partly through higher prices) for the U.S. and world. Slow immigration. Would have small effect on inequality and would retard economic growth given the low birth rate in the U.S. (and other developed countries). Raise minimum wages. Usual criticism by economists is MW lowers employment of least skilled. Evidence shows these effects small. MW may be attractive policy to decrease lefttail inequality; does little to expand middle class. Encourage unionism in the private sector. Unionization will decrease inequality and raise wages for covered workers. But wage increases not fully offset by productivity increases. U.S. unionism associated with poorer firm performance lower profits, investment, growth. Traditional unionism may not be the best way to enhance worker voice and cooperative employment participation in the workplace. 17