Presentation on theme: "Skill Use and Technical Change: International Evidence National Institute of Economic and Social Research Mary OMahony Catherine Robinson Michela Vecchi."— Presentation transcript:
Skill Use and Technical Change: International Evidence National Institute of Economic and Social Research Mary OMahony Catherine Robinson Michela Vecchi (email@example.com)
Previous literature Consider the impact of information and communications technology (ICT) on skill use. Previous literature focused on skill bias technical change. The evidence on balance suggests technical change has been moving in a skill biased direction. and some evidence this is linked to ICT. Focus of previous work has been on comparing highest and lowest skill groups. Misses the dynamics within intermediate groups.
This paper: Examines skill use in the 1980s and 1990s for three countries, the US, the UK and France. The analysis is based on industry data covering all sectors within the non-agricultural market economy. Labour input is divided into five types (six for France) according to education/qualification levels. We also separately identify IT occupational groups, with a division into two qualification levels, those with degrees and others.
Chart 1.a. Trends in labour shares US.
Chart 1.b. Trends in labour shares UK.
Chart 1.c. Trends in labour shares France.
IT labour: occupations
The decomposition approach (Berman, Bound and Griliches, 1994) for i=1, …, n industries, j skill groups. P j is the share of each skill type in total employment, P ji. is are the industry shares of type j skill group, - denotes average across industries. S i is the share of employment in industry i. If within effect dominates then supports SBTC Can do similar decomposition for wage bill shares between within
Chart 2a. Decomposition results, US employment
Chart 2b. Decomposition results, UK employment
Chart 2c. Decomposition results, French employment
Wage share equations Translog cost function: standard wage share (Ws) equation, i industries, j skill types, w = wage rates, K = capital and Y = output, n is the lowest skill group and WT is the total wage bill Wage bill shareSubstitution Capital-skill complementarity New technology
Wage share equations Adoption versus use of new technology (Chun 2003) Use: Share of ICT in total capital services - standard technology variable Adoption: Chun - age of ICT capital This paper: IT specific labour requirements ratio IT workers to ICT capital
Industry data set US: 31 industries, annual 1979-2000 UK: 27 industries, annual 1982-2000 France: 30 industries, annual 1982-2000 Output: Value added Capital: Tornqvist index of capital services, three ICT types (computers, software and communications equipment) and three non-ICT (structures, other equipment, transport equipment)
Regression Results For the US: NB: time dummies were included for all specifications
US results continued…
Results for the UK
UK results continued…
Results for France
French results continued…
Conclusions Nature of skill bias changes across time New technology increased wage premiums for highly skilled workers more in 1980s than 1990s in the US. Evidence that this technology is moving more towards increasing wage premiums for those with intermediate skills over time. Adoption effect highly significant in the US - consistent with findings by Chun (2003) Evidence for the UK is more mixed but there is some suggestion of increasing skill bias for intermediate skill levels.
Conclusions In contrast, in France only one intermediate category shows any positive impact from ICT and the lowest intermediate category shows a negative impact in the 1990s. Probably reflects the differences in the labour market institutions in France High supply of educated workers and their relatively low wage