Title: Uncertainty, inequality, and cheap steel Peter B. Meyer Office of Productivity and Technology U.S. Bureau of Labor Statistics 5 October 2004 T
1Uncertainty, inequality, and cheap
steelPeter B. MeyerOffice of Productivity and
TechnologyU.S. Bureau of Labor Statistics5
October 2004TübingenOutline1.
Technological change, 1866-1881 2. Weeks report
data set3. Measures of earnings inequality4.
Uncertainty vs. skill bias5. Occupations and
institutions
2Iron and steel materials
Bessemer steel and open hearth steel were new to
the U.S. in the 1860s and 1870s. That completed
the technology set for 40 years. Wages are
available from blast furnaces and rolling mills.
Iron ore
Blast furnace
castings
Pig iron (includes 4 carbon)
puddling
wrought (or bar) iron (.1 carbon)
Bessemer steel process (1860s)
Rolling mill
open hearth steel processes (1870s)
Rails for railroads
crucible steel for tools, etc (1.9 carbon)
3Improving technology output rose and prices
fell
Source Historical Statistics of US
4Weeks report
- U.S. Census Bureau surveyed manufacturing plants
in the 1880s - Establishments reported wages by job, usually a
retrospective yearly average - Most observations are for a job title, not a
worker. - 1047 establishments in 48 industries
- 104,413 wage observations but some are piece
rates not day-wages
5Industries in the data
Metal work industries
Iron blast furnaces
Machinery making
Rolling mills
Foundries
Hardware and cutlery
Tin and sheet iron works
Textiles and clothing
Wood work industries
Ag / food / forestry products processing
Mining and minerals processing
Construction materials
6Earnings inequality within metalwork industries
y-axis has coefficient of variation
Inequality rose within the blast furnace and
rolling mills industries starting about 1869, but
did not rise in other metalwork industries
7Earnings inequality within other industries
y-axis has coefficient of variation
Inequality rose within the blast furnace and
rolling mills industries starting about 1869, but
did not rise in the other grouped industries
8Earnings dispersion in iron and steel (from
table 1a)
Variance of weighted log-wages after year fixed effects removed Variance of weighted log-wages after year fixed effects removed
1855-1869 1870-1881
Blast furnaces and rolling mills .198 (N2172) .284 (N7251)
All other manufacturing wages .292 (N31662) .278 (N58614)
Earnings inequality within iron and steel
industries rose.
9Earnings dispersion in iron and steel (from table
1b)
Variance of residuals of mens weighted log-wages after year and job fixed effects regression Variance of residuals of mens weighted log-wages after year and job fixed effects regression
1855-1869 1870-1881
Blast furnaces and rolling mills .090 (N1915) .112 (N5731)
All other manufacturing wages .100 (N27276) .101 (N49932)
Residual inequality within iron and steel
industries rose with job held constant.
10Earnings dispersion in iron and steel (from table
1c)
Variance of residuals weighted log-wages after regression in Table A7 Variance of residuals weighted log-wages after regression in Table A7
1855-1869 1870-1881
Blast furnaces and rolling mills .097 (N2172) .145 (N7251)
All other manufacturing wages .120 (N31662) .121 (N58614)
Residual earnings inequality within iron and
steel industries rose, perhaps after any
regression.
11Uncertainty and/or skill bias
- Skill-biased technological change preexisting
skills are more productive and better rewarded by
working with the new technology - Katz and Murphy (1992) Juhn, Murphy, and Pierce
(1993) Bound and Johnson (1992) Murphy and
Welch (1992)) - Technological uncertainty people and firms
experiment with the new technology and have
divergent results. - Nelson (1961) Malerba (1985) Dosi (1988)
Rosenberg (1996) Greenwood and Yorukoglu (1997)
12Uncertainty model
- Suppose workers, managers, and firms choose
between - doing the work in the standard way
- quitting
- But with technological change, more choices are
possible - try different technology
- reorganize the work process
- invent something new
- More options ? More dispersed outcomes
- Statistical evidence Residuals from a
substantive wage regression rise for iron and
steel after 1869
13Puddlers, rollers, and inside contracting
- Puddlers make wrought iron by craft work
- Rollers send iron or steel through rolls to
press impurities out and to shape the material - Inside contracting skilled craftspeople hired
and organized their own employees, who were paid
as a group by output. - Contracting declined after the 1870s.
14The Rise of Pittsburgh
- Pittsburgh made no Bessemer steel until 1875, but
had advantages - Coal sources are nearby
- Iron ore travels cheaply over Great Lakes
- Much iron production already
- Home of the railroad managers
- Production near Pittsburgh grew
- By 1880 it was a fast growing steel producing
area - Remained dominant for decades
- In the data, workers receive higher wages in
Pittsburgh - a location effect
15New kinds of management
- Carnegies plants were very successful
- They did cost accounting of inputs
- The iron and steel industry learned from the
railroads, which were larger (Pennsylvania
Railroad in 1875 had over 50,000 workers.
Largest single US plant was Cambria Iron with
about 4000.) - Managers take charge of tools
- Chemists hired formal RD begins
- Carnegies plants pay high wages and have few
strikes during 1870-1881 - High wages at Carnegie plant in this data -- a
firm effect on wages
16Managers, contd
- Articles in technical journals about management
first appear in the 1870s - Accounting, journals, consultants, and RD
these are approaches to managing information,
replacing craft work and inside contractors
Wages of managers in iron blast furnaces (Figure
5)
Wages of managers of glass-making establishments
17Conclusion
- Earnings inequality rose in the iron and steel
sector - Dimensions of wage differentiation include
- Some jobs faced change and opportunity
- Certain locations won out
- Managers faced growth, applied cost accounting,
and research and development - Inequality came from opportunity
- novelty
- uncertainty
- experimentation
- information management