The U.S. economy has expanded rapidly in recent years, with total factor productivity (the source of growth most closely identified with technological gains) rising sharply since the mid-1990’s (see e.g., Bureau of Labor Statistics, 1999; William Gullickson and Michael J. Harper, 1999; Mun S. Ho et al., 1999; Daniel E. Sichel, 1999). This strong aggregate performance and the well-documented explosion of investment in computers and other high-tech equipment have led many to believe that the United States has experienced a permanent, technology-led growth revival. It is essential, however, to disaggregate estimates of economic growth to the industry level to understand the new trends in the U.S. economy. Productivity growth, the ability to produce more outputs from the same inputs, differs widely among industries. For the economy as a whole, negative productivity growth in one industry can offset positive productivity growth in another, and Jorgenson (1990) shows that a measure of productivity based solely on aggregate data is valid only under very stringent conditions. We avoid the limitations of an aggregate measure of productivity by decomposing U.S. growth across industries for the period 1958–1996. By breaking down the U.S. economy into 37 industries (35 private industries, private households, and general government), we identify the contribution of each industry to aggregate productivity growth. This enables us to isolate the underlying sources of gains in productivity and provides a better understanding of the forces driving the U.S. economy. Economy-wide productivity from an aggregate production function increased 0.45 percent per year during 1958–1996, while methodology developed by Evsey Domar (1961) for aggregating over industries yields an aggregate estimate of 0.48 percent. Over the same period, however, industry productivity growth ranged from 1.98 percent in Electronic and Electric Equipment to 20.52 percent in Government Enterprises, highlighting fundamental differences in technology and productivity growth across industries. These results show that the aggregate production function provides a reasonable estimate of productivity trends over long periods but also masks important differences among industries.
[1]
Kevin J. Stiroh,et al.
Computers, Productivity, and Input Substitution
,
1998
.
[2]
D. Jorgenson,et al.
The Quality of the U.S. Work Force, 1948-95
,
1999
.
[3]
Daniel E. Sichel,et al.
Computers and Aggregate Economic Growth: An Update
,
1999
.
[4]
Is Embodied Technology the Result of Upstream R&D? Industry-Level Evidence
,
2001
.
[5]
Dale W. Jorgenson,et al.
Productivity and Economic Growth
,
1990
.
[6]
Michael J. Harper,et al.
Possible Measurement Bias in Aggregate Productivity Growth
,
1999
.
[7]
Dale W. Jorgenson,et al.
Productivity and U.S. Economic Growth
,
1999
.
[8]
Kevin J. Stiroh,et al.
Do Computers Make Output Harder to Measure?
,
2001
.
[9]
Robert H. McGuckin,et al.
Computers and Productivity: Are Aggregation Effects Important?
,
2002
.
[10]
E. Dean,et al.
The accuracy of the BLS productivity measures
,
1999
.
[11]
D. Jorgenson,et al.
Japan-U.S. industry-level productivity comparisons, 1960-1979
,
1987
.
[12]
G. Clark.
European Pensions and Global Finance: Continuity or Convergence?
,
2001
.
[13]
C. Corrado,et al.
Decomposition of Productivity and Unit Costs
,
1999
.
[14]
E. Domar,et al.
ON THE MEASUREMENT OF TECHNOLOGICAL CHANGE
,
1961
.
[15]
D. Jorgenson,et al.
Information Technology and Growth
,
1999
.
[16]
Stephen D. Oliner,et al.
Improved estimates of fixed reproducible tangible wealth, 1929-95
,
1997
.