The Effects of Technical Change on Labor Market
Inequalities
Andreas Hornstein (Richmond Fed)
Per Krusell (Princeton University)
Gianluca Violante (New York University)
What do we do in the Handbook chapter?
What do we do in the Handbook chapter?
1. We discuss three decades (1970-2000) of data on technology, productivity, and labor market outcomes
2. We explore the hypothesis that technological change has
determined the observed changes in the labor market (and that there may be a feedback)
• Neoclassical theory: competitive labor market, firms as
production functions
• Beyond neoclassical theory: frictional labor market,
What do we do in the Handbook chapter?
1. We discuss three decades (1970-2000) of data on technology, productivity, and labor market outcomes
2. We explore the hypothesis that technological change has
determined the observed changes in the labor market (and that there may be a feedback)
• Neoclassical theory: competitive labor market, firms as
production functions
• Beyond neoclassical theory: frictional labor market,
organization of firms, unemployment, institutions (e.g. unions)
Methodological Issues
• Distinctive feature of this literature: huge variety of approaches
◦ Neoclassical growth/Schumpeterian growth, McCall
search/Lucas-Prescott islands, Mortensen-Pissarides random matching/directed search, Bewley incomplete
Methodological Issues
• Distinctive feature of this literature: huge variety of approaches
◦ Neoclassical growth/Schumpeterian growth, McCall
search/Lucas-Prescott islands, Mortensen-Pissarides random matching/directed search, Bewley incomplete
markets/Arrow-Debreu with limited commitment
• Why such heterogeneity of frameworks?
1. Young field of research
Methodological Issues
• Distinctive feature of this literature: huge variety of approaches
◦ Neoclassical growth/Schumpeterian growth, McCall
search/Lucas-Prescott islands, Mortensen-Pissarides random matching/directed search, Bewley incomplete
markets/Arrow-Debreu with limited commitment
• Why such heterogeneity of frameworks?
1. Young field of research
2. Many possible departures from competitive model
• Challenge for the exposition... we gave priority to presenting a
College-High School Wage Premium
1965 1970 1975 1980 1985 1990 1995 2000
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 Year N o rm a lize d ra tios (1 98 2 =1)
Dynamics of Relative Prices and Quantity of Skills in the U.S. (1963−2002)
Wage Premium COLG−HSG Relative Proportion COLG−HSG
Wage Premium: 1.47
Rise in the Educational Premium I
KATZ-MURPHY (1992)
• From CES production function in skilled and unskilled labor:
log
wst
wut
= σlog
Ast
Aut
+ (1 − σ)log
lut
lst
Rise in the Educational Premium I
KATZ-MURPHY (1992)
• From CES production function in skilled and unskilled labor:
log
wst
wut
= σlog
Ast
Aut
+ (1 − σ)log
lut
lst
◦ σ = 0.3 ⇒ elasticity of substitution around 1.4
◦ relative supply of skills rose at 3% per year
Rise in the Educational Premium I
KATZ-MURPHY (1992)
• From CES production function in skilled and unskilled labor:
log wst wut = σlog Ast Aut
+ (1 − σ)log
lut
lst
◦ σ = 0.3 ⇒ elasticity of substitution around 1.4
◦ relative supply of skills rose at 3% per year
◦ skill-biased technical change (SBTC) grew at 10% per year
• Limits of the pure SBTC hypothesis:
◦ it’s based on an unobservable residual
◦ growth in relative productivity of skilled labor huge, plausible?
Rise in the Educational Premium II
GRILICHES (1969), HOLMES-MITCHELL (2004), JOVANOVIC (1998),
KRUSELL-OHANIAN-RIOS RULL-VIOLANTE (2000)
• Capital-skill complementarity hypothesis:
◦ Capital equipment and skilled labor are complements in
production
◦ Rapid fall of relative price of equipment, due to fast
productivity improvements in ICT, increased the demand for capital in production
• It successfully accounts for the rise in the skill premium,
essentially without the help of growth in the unobserved relative
Further Evidence on K-S Complementarity
FLUG-HERCOWITZ (2000), GOLDIN-KATZ (1998), LINDQUIST (2003, 2004), COEN
PIRANI-CASTRO (2004)
• Large effect of equipment investments on relative wages in a
cross-section of countries
• The theory works also when applied to the evolution of skill premia
in Sweden, where institutions play a big role in the labor market
• It helps explaining cyclical behavior of the skill premium at
business-cycle frequencies
Historical Role of K-S Complementarity
1930 1940 1950 1960 1970 1980 1990 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 Year
Normalized ratios (1929=1)
Dynamics of Relative Prices of Capital and Returns to Education in the U.S. (1929−1995)
Relative Price of Capital Return to College Return to High−School
1930 1940 1950 1960 1970 1980 1990 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 Year
Normalized ratios (1929=1)
Dynamics of Relative Prices of Capital and Returns to Education in the U.S. (1929−1995)
Relative Price of Capital Return to College Return to High−School
1929 = 0.10
Endogenous Skill-Bias
SAMUELSON (1965), ACEMOGLU (1998, 2002, 2003), KILEY (1999),
COZZI-IMPULLITTI (2004)
• If R-D can be directed towards productivity improvements of
Endogenous Skill-Bias
SAMUELSON (1965), ACEMOGLU (1998, 2002, 2003), KILEY (1999),
COZZI-IMPULLITTI (2004)
• If R-D can be directed towards productivity improvements of
different inputs, it will be biased towards the one with the highest return for the innovators
• Sources of SBTC:
1. market size effect:
◦ Exogenous rise in the supply of skilled labor in 1970s
◦ Technology policy: U.S. government shifted expenditures
towards IT goods
2. relative price effect:
Rise in the Returns to Ability
CASELLI (1999), GALOR-MOAV (2000), GALOR-TSIDDON (1997),
GREENWOOD-YORUKOGLU (1997), KRUEGER-KUMAR (2004), LLOYD-ELLIS (1999)
• Nelson-Phelps (1966) Hypothesis: workers endowed with more
innate skills cope better with technological transformations
• Various versions: more able workers...
◦ ... are more productive with the new technology
◦ ... acquire skills specific to the new technology more cheaply
◦ ... are less subject to obsolescence of human capital due to
Returns to Experience by Educational Group
60 65 70 75 80 85 90 95
0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65
Male high school graduates
0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65
Rise in the Experience Premium
AGHION-HOWITT-VIOLANTE (2002), HECKMAN, LOCHNER AND TABER (1998),
WEINBERG (1999)
• General Purpose Technology : transferability of knowledge
Rise in the Experience Premium
AGHION-HOWITT-VIOLANTE (2002), HECKMAN, LOCHNER AND TABER (1998),
WEINBERG (1999)
• General Purpose Technology : transferability of knowledge
improves with GPT, which increases the value of experience and the returns to experience
• Experience facilitates adoption for the low-educated: experience
Age-Profiles of PC Adoption
0 5 10 15 20 25 30 35 40 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
Years of Experience
Fr acti on of W orke rs
Adoption Rate of Computers by High−School Graduates
1984 1987 1993 1997
0 5 10 15 20 25 30 35 40 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 F ract io n of Wor ke rs
Adoption Rate of Computers by College Graduates
Deunionization and the Rise in Inequality
0.18 0.20 0.22 0.24 0.26 0.28 0.30 0.32 2.0 2.5 3.0 3.5 4.066 68 70 72 74 76 78 80 82 84 86 88 90 92
unionization inequality u n io n d e n s it y 9 0 -1 0 w a g e d if fe re n ti a l
Deunionization
ACEMOGLU-AGHION-VIOLANTE (2002), ORTIGUEIRA (2004)
• Traditional view: deunionization is an alternative explanation
• Alternative view: deunionization is caused by technology
◦ Model the union as a coalition of workers heterogeneous in
their skill level, with endogenous participation decision:
◦ Benefit: rent extraction
Deunionization
ACEMOGLU-AGHION-VIOLANTE (2002), ORTIGUEIRA (2004)
• Traditional view: deunionization is an alternative explanation
• Alternative view: deunionization is caused by technology
◦ Model the union as a coalition of workers heterogeneous in
their skill level, with endogenous participation decision:
◦ Benefit: rent extraction
◦ Costs: wage compression for skilled workers, operating cost
• SBTC increases the relative productivity of skilled workers and the
value of opting out of the union to work at the competitive wage
Rise in the Returns to “Luck”
ACEMOGLU (2002), MANUELLI (2000), SHI (2002), VIOLANTE (2002), WONG (2003)
• Models with frictional labor markets, random matching, vintage
Rise in the Returns to “Luck”
ACEMOGLU (2002), MANUELLI (2000), SHI (2002), VIOLANTE (2002), WONG (2003)
• Models with frictional labor markets, random matching, vintage
capital, and Nash bargaining
• An acceleration in the rate of capital-embodied technical change
has two effects:
1. can increase the dispersion of productivities of technologies in operation ⇒ higher wage inequality, since wages depend on
Rise in the Returns to “Luck”
ACEMOGLU (2002), MANUELLI (2000), SHI (2002), VIOLANTE (2002), WONG (2003)
• Models with frictional labor markets, random matching, vintage
capital, and Nash bargaining
• An acceleration in the rate of capital-embodied technical change
has two effects:
1. can increase the dispersion of productivities of technologies in operation ⇒ higher wage inequality, since wages depend on
productivity of machines
2. with vintage human capital, it can increase the dispersion of skills ⇒ higher wage inequality, since wages depend on
Europe vs US: Unemployment and Labor Share
1960 1965 1970 1975 1980 1985 1990 19952 3 4 5 6 7 8 9 10 11 Year Percentage
"Standardized" Unemployment Rate
United States Europe (Average)
Europe vs United States
ACEMOGLU (2003), BERTOLA-ICHINO (1995), CABALLERO-HAMMOUR (1998),
HORNSTEIN-KRUSELL-VIOLANTE (2003), LJUNGQVIST-SARGENT (1998),
MARIMON-ZILIBOTTI (1999), MORTENSEN-PISSARIDES (1998), ROGERSON (2004)
• Krugman’s (1994) hypothesis: similar macroeconomic shocks,
Europe vs United States
ACEMOGLU (2003), BERTOLA-ICHINO (1995), CABALLERO-HAMMOUR (1998),
HORNSTEIN-KRUSELL-VIOLANTE (2003), LJUNGQVIST-SARGENT (1998),
MARIMON-ZILIBOTTI (1999), MORTENSEN-PISSARIDES (1998), ROGERSON (2004)
• Krugman’s (1994) hypothesis: similar macroeconomic shocks,
interacted with different institutions lead to different outcomes
◦ Labor supply view: technology shock ⇒ skill obsolescence
reduces earning power ⇒ jobless workers prefer collecting UI
Europe vs United States
ACEMOGLU (2003), BERTOLA-ICHINO (1995), CABALLERO-HAMMOUR (1998),
HORNSTEIN-KRUSELL-VIOLANTE (2003), LJUNGQVIST-SARGENT (1998),
MARIMON-ZILIBOTTI (1999), MORTENSEN-PISSARIDES (1998), ROGERSON (2004)
• Krugman’s (1994) hypothesis: similar macroeconomic shocks,
interacted with different institutions lead to different outcomes
◦ Labor supply view: technology shock ⇒ skill obsolescence
reduces earning power ⇒ jobless workers prefer collecting UI
to working at the low wage ⇒ u rises
◦ Labor demand view: K-embodied technology shock ⇒ labor
becomes too expensive relative to capital due to wage rigidity
Europe vs United States
ACEMOGLU (2003), BERTOLA-ICHINO (1995), CABALLERO-HAMMOUR (1998),
HORNSTEIN-KRUSELL-VIOLANTE (2003), LJUNGQVIST-SARGENT (1998),
MARIMON-ZILIBOTTI (1999), MORTENSEN-PISSARIDES (1998), ROGERSON (2004)
• Krugman’s (1994) hypothesis: similar macroeconomic shocks,
interacted with different institutions lead to different outcomes
◦ Labor supply view: technology shock ⇒ skill obsolescence
reduces earning power ⇒ jobless workers prefer collecting UI
to working at the low wage ⇒ u rises
◦ Labor demand view: K-embodied technology shock ⇒ labor
becomes too expensive relative to capital due to wage rigidity
⇒ capital substitutes labor in production
◦ Sectoral-transformation view: if one focuses on e (rather
than on u), the US-EU trends start diverging in the 1960s ⇒
Policy Implications
HECKMAN-CARNEIRO (2004), KRUEGER (2004)
• To the extent that the impact of technological change on inequality
Policy Implications
HECKMAN-CARNEIRO (2004), KRUEGER (2004)
• To the extent that the impact of technological change on inequality
is “socially inefficient”, we can speculate on the best policy
• Technology-skill complementarity
Policy Implications
HECKMAN-CARNEIRO (2004), KRUEGER (2004)
• To the extent that the impact of technological change on inequality
is “socially inefficient”, we can speculate on the best policy
• Technology-skill complementarity
◦ important to rebuild (obsolete) skills through training policies
• Technology-ability complementarity
◦ interventions should occur early in the life-cycle
◦ adult training is ineffective, better subsidizing wages for the