Transcript Labor and Multifactor Productivity Growth Formulas

Technological Change and
Economic Growth: the
Interwar Years and the
1990s
Alexander J. Field
[email protected]
All Ohio Eocnomic History Seminar
Ohio State Universiyty
April 30, 2004
The Most Technologically Progressive
Decade of the Century




Field, American Economic Review (2003)
Hint #1: It’s not what you think …
Hint #2: It wasn’t the 1990s...
Hint #3: It wasn’t the 1920s…
The Main Argument

The years 1929-1941 were, in the
aggregate, the most technologically
progressive of any comparable
period in U.S. economic history.
Labor and Multifactor Productivity
Growth Formulas
Y = real output
N= labor hours
K=capital input
Y/N = Labor Productivity
y – n = Labor Productivity growth
(lower case letters = compound annual average rates of growth)
Y = A KβN1-β = Production function (Cobb Douglas, crts)
A = Y/(KβN1-β) = Multifactor Productivity
a = y – βk – (1-β)n = Growth Rate of MFP
y - n = a + β (k - n) = Growth rate of labor productivity
Disclaimers - 1

What we measure in the residual contains not
only serendipitous or accidental discovery of
useful knowledge, but a variety of related
influences, including, but not limited to:
• The outcomes of focused research and development
activities
• The influence of scientific and educational infrastructures
• Economies of scale and network effects
• Learning by doing
• Reallocation of economic activity from sectors with low
to higher value added per worker
• New organizational blueprints or managerial practices
Disclaimers - 2

Under certain conditions, the residual will
underestimate the impact of technological change
because of linkages between such change and
the rate of saving and capital accumulation
• Biased technical change that increases the return to
capital or to highly educated labor may shift income to
households with higher saving propensities, increasing
the aggregate saving rate
• Technical change that increases the real after tax return
to capital may elicit larger flows of saving
• Innovations that reduce the relative price of capital
goods will, for a given saving rate measured at base
period prices, be associated with a larger real volume of
investment
Abramovitz – David (1999)
CAAGR of MFP
1890-1905
1905-29
1929-48
1948-66
1966-89
1.28
1.38
1.54
1.31
.04
Gordon (2000b)
CAAGR of MFP
1870-1891
1891-1913
1913-1928
1928-1950
1950-1964
1964-1972
1972-1979
1979-1988
1988-1996
.39
1.14
1.42
1.90
1.47
.89
.16
.59
.79
Conventional Wisdom
The measured peak in MFP growth
rates between 1929 and 1948 is
principally the consequence of the
production experience of World War
II: a persisting benefit of the
enormous cumulated output as well
perhaps of spinoffs from war related
R and D.
New Argument



Peak MFP growth between 1929 and 1948 is
primarily attributable to an exceptional
concatenation of technical and organizational
advances across a broad frontier of the American
economy prior to full scale war mobilization
Governmental and university funded research, as
well as the maturing of a privately funded R and
D system that began with Edison at Menlo Park
played a role
So too in some sectors, such as railroads, did the
“kick in the pants” of cut off of easy credit
availability and declines in demand
Why do we credit World War 2 with
establishing the foundations for
postwar prosperity?


Sheer volume of output between 1942 and 1945 is exceptional
Remarkable successes in such sectors as airframes and
shipbuilding
• Between 1942:1 and 1944:4 airframe production
increased by a factor of six, and labor productivity grew by 160
percent. This is a compound annual average growth rate of output per
hour of 34.7 percent
In shipbuilding: in one ten-month period alone, the
number of hours required to build a Victory ship
fell by half (U.S. Bureau of Labor Statistics,
1946, pp. 897–98). On an annualized basis, this is a growth in output
per hour of 83 percent a year
Why we should be skeptical





Overall increase in labor productivity in munitions sector
between 1939 and 1945 was 25 percent – far below
standout sectors (Brackman and Gainsbrugh, 1949). This
is a growth in output per hour of 3.71 percent per year,
respectable, but hardly surprising given the more than $10
billion of public sector capital invested in the defense sector
Short period of full scale war production: roughly three and
a half years
Spillovers – which direction?
War drained skilled labor, managers, and capital from
civilian sector. Output per hour stagnated in 1942-44 in the
civilian sector
Swollen productivity numbers are partly the result of a
temporary shift of output to sectors with traditionally high
ratios of value added per worker – could not persist
Why 1941?





1937 – unemployment 14.3 percent
1940 - unemployment even higher – 14.6
percent
1941 unemployment is 9.9 percent (6
percent according to Darby)
Only 2.5 percent of cumulated war
spending 1941-1945 had been undertaken
by the end of 1941
1941 is the closest we come to recovery
before full scale war mobilization
United States, Private Non-Farm Economy
CAAGR of MFP, 1919-1948
1919-1929
1929-1941
1941-1948
Solow
Kendrick
.78%
2.36%
.89%
2.02%
2.31%
1.29%
Solow (1957)

“there does seem to be a break at
about 1930. There is some evidence
that the average rate of progress in
the years 1909–29 was smaller than
that from 1930–49” (Solow, 1957, p.
316)
Kuznets and War Planning



Simon Kuznets needed to estimate the potential
output of the U.S. economy to determine war
production plans consistent with planned force
levels and civilian consumption.
His estimates came in considerably higher than
most had expected, leading the military to
multiply their production targets and forcing
Kuznets and others to fight a rear guard action to
bring them down to a realistic level
The outward shift of the production possibility
frontier during the Depression years -- largely
unrecognized until then -- was the principal
reason potential output in 1942 was so much
higher than had been anticipated.
Sectoral Evidence


R and D employment in manufacturing
Time path of key innovations
• Kleinknecht
• Schmookler
• Mensch


MFP growth in telephone, railroads,
electric utilities
Build out of surface road system: network
effects – impact on productivity growth in
trucking and warehousing
R and D employment in US
Manufacturing



1927:
1933:
1940:
6,274
10,918
27,777
Source: National Research Council data;
Mowery and Rosenberg, 2000
30
25
20
15
10
5
0
1960-69
1950-59
1940-49
1930-39
1920-29
1910-19
1900-09
1890-99
1880-89
1870-79
1860-69
1850-59
Kleinknecht (1987)
Product
Process
Instrument
Other
Total
Basic Innovations, 1900-1959
Schmookler
Mensch
70
60
50
40
30
20
10
0
1955-1959
1950-1954
1945-1949
1940-1944
1935-1939
1930-1934
1925-1929
1920-1924
1915-1919
1910-1914
1905-1909
1900-1904
The Irony of Secular Stagnation



At precisely the moment when Alvin Hansen and
others were developing theories of secular
stagnation, the U.S. economy was experiencing
its greatest technological efflorescence, a period
of creativity which, in the aggregate, remains
unmatched to this day.
His Harvard colleague, Joseph Schumpeter had a
better fix on what was going on, although he
misjudged the terrain on the road to socialism.
Schumpeter’s homage to “creative destruction”
was developed against the backdrop of what in
fact has turned out to be the most technologically
dynamic epoch of the twentieth century.
CAAGR of MFP, 1919-1948
Telephone Electric
Utilities
1919-1929 1.60%
1929-1941 2.01%
1941-1948 0.53%
2.51%
5.55%
5.87%
Railroads
1.63%
2.91%
2.56%
Table 4: Compound Annual Average Growth Rates of Net Stock of Street
and Highway Capital, United States, 1925-2000





1925-1929
1929-1941
1941-1948
1948-1973
1973-2000
6.00%
4.32%
0.08%
4.15%
1.63%
Street/Highway Capital as % of
Private Fixed Capital
Street/Hway K Private Fixed K





1929 $16,415 $253,987
1941 $30,861 $289,487
1948 $47,892 $582,248
1973 $290,389 $2,698,194
2000 $1,423,833 $21,464,786
6.46%
10.66%
8.22%
10.76%
6.63%
Two Waves of Government
Investment: the 1930 and the
1940s






Both decades experienced substantial
government investment in physical capital
1930s: Infrastructure
1940s: Over $10 billion of GOPO capital in
manufacturing, much of it equipment, especially
machine tools
The 1930s investment generated spillovers that
augmented PNE MFP growth, particularly in
trucking and warehousing, and, through
complementarities, in railroads.
The 1940s investment was associated with
declining MFP growth in manufacturing and for
the economy as a whole.
What does this imply for the generality of the
equipment hypothesis?
Summary – Field (2003)


The 1930s were characterized by productivity
advance across a broad frontier of the U.S.
economy. The expansion of potential output
between 1919 and 1941 laid the foundation for
postwar prosperity, at the same time that it
enabled successful prosecution of the war.
In the light of this finding we need to rethink our
understanding of the defining contours and
determinants of U.S. economic growth in the
twentieth century.
Field (2004)



Compares 1929-41 with 1919-29 on the
one hand, and 1995-2000 on the other
Disaggregates, by broad sector,
contributions to MFP growth in these
different periods
Reassesses IT’s impact on economic
growth in the 1990s, as well as the
broader utility of the GPT concept with
which it is closely associated.
Main Arguments




MFP Growth in the 1920s was almost entirely a
story about manufacturing
In the 1930s, manufacturing’s contribution
declined, although remaining high in absolute
terms. Transport and Public Utilities played a
much more important role. The same was true
to a lesser extent of distribution.
Although MFP growth between 1995 and 2000
was triple what it had been during 1973-95, it
was less than half what it was between 1929 and
1941.
The IT revolution was responsible for about two
thirds of MFP growth, and about a third of labor
productivity growth between 1995 and 2000
CAAGR of MFP, PNE,
United States, 1919-2000
1919-1929
1929-1941
1941-1948
1948-1973
1973-1989
1989-2000
2.02
2.31
1.29
1.90
.34
.78
1973-1995
1995-2000
.38
1.14
Sources:
1919-48: Field (2003); Kendrick (1961)
1948-2000: Bureau of Labor Statistics: www.bls.gov
CAAGR of Labor Productivity,
United States, 1919-2000
1919-1929
1929-1941
1941-1948
1948-1973
1973-1989
1989-2000
2.27
2.35
1.71
2.88
1.33
1.97
1973-1995
1995-2000
1.40
2.43
Sources:
1919-48: Kendrick (1961), Table A-23.
1948-2000: Bureau of Labor Statistics: www.bls.gov
Labor Productivity Growth



Roughly comparable over 1919-29, 192941, and 1995-2000.
The 1930s were exceptional because
advance took place in the virtual absence
of capital deepening
The 1948-73 period remains the golden
age of living standard improvement,
because of the combined effects of
respectable MFP growth and robust rates
of capital deepening.
Labor Quality - 1





How much of the growth in output per hour between 192941 is attributable to labor quality improvement?
Margo (1991, 1993) has emphasized selective retention of
higher quality workers as employment drops in a recession
Not relevant for peak to peak comparisons; these
composition effects would have been unwound as economy
returned to full employment
1941 is much closer to full employment than 1940 or 1937,
but still had 9.9 percent unemployment, vs. less than 4
percent in 1929
In retrospect it would have been helpful for my research
had Japan delayed attack on Pearl Harbor for another 8 to
12 months, so the U.S. economy could have continued its
then rapid movement toward full employment before full
scale war mobilization took place
Labor Quality 2





Goldin (1998) has emphasized rapid rise in high
school graduation rates in the 1930s.
R and D employment data for manufacturing and
Margo’s analyses reflect strong demand for
managerial, scientific, and technical personnel
during this period
Opportunity cost of high school attendance
dropped as probability of a non high school grad
being unemployed rose
Build out of surface road network facilitated high
school attendance
Influx of human capital fleeing Hitler’s Europe
Labor Quality 3



Labor Quality did rise between 1929 and 1941, but changes
in labor supply were probably not a major or dominant
influence on growth in output per hour
If we calculate rate of growth of output per hour between
1929-41 using adjusted hours, where adjusted hours take
into account labor quality improvement, increase in output
per hour is 6 percent less
If selective retention were the dominant influence on
growth as we came out of recession, we should have had a
larger increment to output per hour moving from 19.1 to
14.6 percent unemployment between 1938 and 1940 than
we did moving from 14.6 to 9.9 percent unemployment
1940 to 1941. But the reverse was true.
1929-41 Manufacturing MFP
Calculation





Output: 3.81 percent/year
Hours: 1.35 percent/year
Capital: .85 percent/year
MFP:
2.60 percent/year
Hours and output data from Kendrick; capital input data
from BEA Fixed Asset Table 4.2
1941-48 Manufacturing MFP
Calculation





Output: 2.20 percent/year
Hours: 2.17 percent/year
Capital: 4.02 percent/year
MFP:
-.52 percent/year
Hours and output data from Kendrick; capital input data
from BEA Fixed Asset Table 4.2
CAAGR of MFP, Manufacturing,
United States, 1919-2000
1919-1929
1929-1941
1941-1948
1948-1973
1973-1995
1995-2000
5.12
2.60
- .87
1.52
.66
2.09
Sources: Field (2004), Kendrick, (1961); Bureau of Economic Analysis Fixed
Asset Table 4.2; Bureau of Labor Statistics, Series MPU300003 (B).
Manufacturing, 1919-1929

Share of PNE (1929):
.333
CAAGR, MFP(1919-29):
5.12%
Cont to PNE MFP Growth:
1.71%
PNE MFP Growth (1919-29): 2.01%

Manufacturing’s Share:



85%
MFP Growth, Durable Manufacturing
19191929
Manufacturing
Durable goods..................................
5.12
5.06
19291948
19491973
19732000
19731995
19952000
1.71
1.52
0.93
0.66
2.09
1.43
1.48
1.57
1.19
3.27
Lumber and wood products.....................
2.49
1.42
1.67
0.65
0.98
-0.78
Furniture and fixtures.......................
4.14
2.00
0.60
0.74
0.69
0.96
Stone, clay, and glass products..............
5.57
2.09
1.09
0.51
0.49
0.57
Primary metal industries.....................
5.36
1.30
0.39
-0.12
-0.36
0.95
Fabricated metal products....................
4.51
1.36
0.54
0.20
0.19
0.26
Industrial, Commercial Machinery
2.82
1.62
0.71
2.93
2.31
5.65
Electric and electronic equipment............
3.45
2.55
2.07
3.85
3.09
7.18
Transport Equipment
8.07
0.39
1.47
0.18
0.05
0.72
4.47*
2.36*
1.75
0.97
1.07
0.54
1.55
0.47
0.33
1.10
Instruments and related products.............
Miscellaneous manufacturing .......
MFP Growth, NonDurable Manufacturing
1919-29
Nondurable goods...............................
4.89
1929-48
2.22
1949-73 1973-2000 1973-95 1995-2000
1.32
0.14
0.06
0.47
Food and kindred products....................
5.18
1.47
0.67
0.26
0.39
-0.31
Tobacco products.............................
4.28
4.19
-0.62
-3.71
-2.92
-7.19
Textile mill products........................
2.90
3.28
2.29
2.23
2.32
1.87
Apparel and other textile products...........
3.90
0.63
0.72
1.00
0.91
1.41
Paper and allied products....................
4.54
2.33
1.58
-0.11
-0.31
0.74
Printing and publishing......................
3.67
1.43
0.48
-0.73
-0.94
0.22
Chemicals and allied products................
7.15
3.36
2.51
-0.25
-0.42
0.51
Petroleum and coal products..................
8.23
1.70
0.82
0.03
-0.14
0.77
Rubber and miscellaneous plastics products...
7.40
2.07
0.96
0.64
0.42
1.60
Leather and leather products.................
2.88
1.71
0.02
0.58
0.21
2.23
Manufacturing, 1929-1941





Share of PNE (1941):
.426
CAAGR, MFP (1929-41):
2.60%
Cont to PNE MFP Growth:
1.11%
PNE MFP Growth (1929-41): 2.31%
Manufacturing’s Share:
48%
Manufacturing, 1995-2000





Share of PNE (2000):
.214
CAAGR, MFP(1995-00):
2.08%
Cont to PNE MFP Growth:
.45%
PNE MFP Growth (1995-00): 1.14%
Manufacturing’s Share:
39%
Transport and Public Utilities 19191929





Share of PNE (1929):
.14
CAAGR, MFP(1919-29):
1.86%
Cont to PNE MFP Growth:
.27%
PNE MFP Growth (1919-29): 2.01%
Sector’s Share:
13%
Transport and Public Utilities 19291941





Share of PNE (1941):
.123
CAAGR, MFP(1929-41):
4.67%
Cont to PNE MFP Growth:
.58%
PNE MFP Growth (1929-41): 2.31%
Sector’s Share:
25%
Transport and Public Utilities 19291941



26 percent of sector MFP growth
comes from railroads
40 percent of sector MFP growth
comes from trucking & warehousing
10 percent of total private non farm
MFP growth comes from trucking &
warehousing
Table 7
MFP Growth, Transportation and Public Utilities, 1929-1941
1929-41
Share of Share Share of Subsector
NI*100
of
Covered
MFP
1941(1948) T & PU Subsectors Growth
Railroad transportation......................
3.27 0.372
Local and interurban passenger transit.. 0.66 0.075
Trucking and warehousing..................... 1.03 0.117
Water transportation.........................
0.42 0.048
Transportation by air........................
0.18 0.021
Pipelines, except natural gas................ 0.10 0.012
Transportation services...................... 0.13 0.014
Telephone and telegraph..................... 1.30 0.148
Radio and television.........................
0.10 0.011
Electric, gas, and sanitary services...
1.61 0.183
8.81
TOTAL
Subsector
Contribution
MFP Growth
0.420 2.91
0.084 3.02
0.132 13.57
0.054 1.47
0.023 14.69
0.013 4.48
1.22
0.25
1.80
0.08
0.34
0.06
0.167 2.02
0.34
0.104
0.57
5.55
4.67
Wholesale and Retail Trade
1919-1929





Share of PNE (1929):
.201
CAAGR, MFP(1919-29):
.80%
Cont to PNE MFP Growth:
.17%
PNE MFP Growth (1919-29): 2.01%
Trade’s Share MFP Growth:
8%
Wholesale and Retail Trade
1929-1941





Share of PNE (1941):
.223
CAAGR, MFP(1929-41):
1.81%
Cont to PNE MFP Growth:
.40%
PNE MFP Growth (1929-41): 2.31%
Trade’s Share MFP Growth:
18%
Wholesale and Retail Trade
1995-2000





Share of PNE (2000):
.223
CAAGR, MFP(1995-00):
.70 %
Cont to PNE MFP Growth:
.16%
PNE MFP Growth (1995-00): 1.14%
Trade’s Share MFP Growth:
14 %
Let’s Remove Religion from the
Analysis of Productivity Trends






Late 1990s: New economy skeptics and true believers
The move from skeptic to believer as a matter of “Getting
Religion”; akin to a process of spiritual conversion
Use of vignettes and anecdotes
Is this really the way to discipline our conclusions with
data?
Focus on what the data actually reveal, rather than what
they might reveal in the future, or we hope they will reveal
in the future
With the marked acceleration in both MFP and labor
productivity growth 1995-2000, one can no longer claim
the statistical apparatus is incapable of picking up effects of
IT.
Reckoning The Impact of IT on
Labor Productivity Growth
Conventional Framework: 3 parts
1. MFP growth within IT producing sector
2. MFP growth within IT using sectors
(spillovers)
3. That portion of the impact of capital
deepening on labor productivity
growth associated with the
accumulation of IT capital goods
Why Include the Third Component?






In the absence of IT, saving flows would have been
congealed in a slightly inferior range of capital goods.
Social Savings Analogy. Debate between Rostow on the
one hand and Fishlow and Fogel on the other. Fogel and
Fishlow imagined worlds in all respects similar save the
availability of railway technology.
Fogel: Because of the railroad, GDP was higher in 1890.
But not that much higher: the availability of the technology
made a difference of 4 percent
Over a 25 year period, what kind of an annual increment to
MFP growth does one need to produce a 4 percent increase
in GDP?
.15 percentage points
In MFP terms, that’s the railroad’s contribution
Rostow Redux

Would it be reasonable for supporters of
the “indispensability” thesis to object that
Fogel’s estimate vastly underestimated the
contribution of the railroad to labor
productivity growth and US standards of
living because it failed to take into account
that portion of labor productivity growth
attributable to the (substantial) fraction of
capital deepening associated with the
accumulation of railway permanent way,
bridges, tunnels, roundtables, stations,
locomotives, and rolling stock?
Solow Again


It remains worthwhile distinguishing
between the effects on labor
productivity of thrift and innovation
There is little evidence that saving
rates accelerated in the presence of
the IT revolution
Sources and Uses of Saving
S = I + (G-T) + (X-M)
Private Domestic Saving funds Private Domestic Capital Accumulation, the
Government Deficit, and capital accumulation outside of the Country
I = S + (T-G) + (M-X)
Private Domestic Capital Accumulation must be funded by the sum of
private domestic saving, government saving, and the capital account
surplus (inflows of foreign saving).
End of Century Tangible Capital
Accumulation





Whatever one can say about the rising importance of
human capital formation in the twentieth century, a
distinguishing feature of the 1995-2000 was a rise in
tangible capital accumulation, with a heavy emphasis on
short lived equipment.
Capital services growth accelerated from 3.94 percent per
year (1973-95) to 5.38 percent per year (1995-2000).
The capital deepening that made this possible required
capital accumulation that had to be financed.
As growth of IT capital services accelerated from .41 to
1.03 percent per year, growth of non IT capital services
essentially halted, dropping from .30 to .06 percent per
year.
Inflow of foreign saving represented a slowing of capital
services growth outside of the country.
Sources and Uses of Saving, US, 1995-2000
1995
2000
1,143.8
238.2
1,755.4
319.8
Personal Saving
302.4
201.5
-100.9
Retained Business Earnings
Depreciation Allowances
Gross Business Saving*
203.6
743.6
963.6
152.6
1,017.9
1,170.5
206.9
Gross Government Saving
-8.5
435.8
444.3
Capital Account Surplus
98.0
395.8
297.8
Statistical Discrepancy
26.5
-128.5
-155.0
Gross Private Domestic Investment
Gross Government Investment
TOTALS
Change
in uses
Change
in
sources
611.6
81.6
693.2
693.1
Table 13
MFP Contribution to Labor Productivity Growth and Acceleration, 1995-2000
Labor Productivity Growth, 1995-2000a
2.46
MFP a
1.14
Capital deepening a
1.05
Labor Composition a
.26
Labor Productivity Growth Acceleration, 1995-2000 vs. 1973-95b
1.09
MFP b
.76
Capital Deepening b
.32
Labor Composition b
-.01
a
percent per year
percentage points
Note: Components do not sum exactly to aggregates due to rounding
errors.
b
Table 16
Sectoral Contributions to MFP growth, 1995-2000
Share of PNE
Sectoral MFP
Growth
Contribution
to PNE MFP
Growth
Share of PNE
MFP Growth
Manufacturing
.214
2.08
.45 .39
Trade
.223
.70
.16 .14
Other
.563
.94
.53 .47
TOTAL
1.00
1.14
Sources: Sectoral Shares: see Table 3
Note: Private Nonfarm economy excludes nonfarm housing, health, agriculture, and government, which leaves 72.4 percent of value added.
MFP growth Manufacturing: see Table 5
MFP growth Trade: see text
IT’s Contribution




Credit IT with all MFP growth in
Manufacturing
Credit IT with one third of MFP growth in
the rest of the economy
.68 of 1.14 percent per year (about 60
percent) of MFP growth between 1995 and
2000 attributable to IT
28 percent of labor productivity growth
1995-2000 is attributable to the IT
revolution (.68/2.46 percent per year)
The Equipment Hypothesis

De Long and Summers (1991)
• Cross country regressions using 1960-85 data show a
relationship between share of equipment investment in
GDP and the growth of output per worker

Auerbach, Hasset, and Oliner (1994)
• Criticize econometrics, question whether conclusion is
applicable to US

Field (2004)
• The drop in relative prices of equipment, and increase in
equipment’s share of capital formation in the US
coincides with a long term decline in the rate of MFP
advance from its high rates in the interwar period to its
virtual disappearance in the last quarter of the century
• Is 1995-2000 an aberration, or a fundamental turning
point in US. productivity history?
Multifactor Productivity Growth Rate, Private Non-Farm
Economy, United States, 1919-2000
2.5
2
1.5
Series1
1
0.5
0
1919-1929
1929-1941
1941-1948
1948-1973
1973-1989
1989-2000
GPT’s





Distinguish between the proposition that it sometimes takes
a long time for the productivity benefits of new
technological complexes to be reaped, and the usefulness
of the concept of a GPT
Potential multiplicity of candidates: Steam, electricity, IT
are most frequently identified, but chemical engineering,
the internal combustion engine, radio transmission and the
assembly line have also been mentioned.
Identification of one or several GPT’s often offers an
appealing narrative hook, but the criteria for designating
them are not universally agreed upon, in spite of continuing
efforts to nail them down.
Why isn’t the railroad also a GPT?
Gordon (2004) suggests that use by both households and
industry is a criterion. This works for electricity and the
internal combustion engine. But steam?
GPTs

Bessemer and Siemens Martin processes were
industry specific, and would clearly not pass
muster as GPTs. They offered, to use David’s
words, “complete, self-contained and
immediately applicable solutions.” This was not
the case for the product whose production they
enabled. Does that make steel a GPT? It took
Carnegie and others time to persuade users they
should make skyscrapers, plate ships, and
replace rails with it. Cheap steel in turn
encouraged complementary innovations such as,
in the case of taller buildings, elevators.
GPTs


If one follows the impact of product and process
innovations far enough through the input-output table, one
will eventually find products or technological complexes
used as inputs in many other sectors, with the potential to
generate spillover effects in using sectors. These
processes, products, or complexes are the consequence of
many separate breakthroughs as well as learning by doing,
much of which has been sector specific. IT for example,
has required advances in sector specific semiconductor
manufacturing, and the thin film technology and
mechanical engineering that underlies most mass storage,
let alone software.
Because of the potential for multiplying GPT candidacies
and the lack of an authoritative tribunal applying uniform
rules passing judgment about which ones qualify, economic
and technological history may be better off without the
concept.
Conclusions

IT was responsible for about 60
percent of MFP growth and about 28
percent of labor productivity growth
between 1995 and 2000
Conclusions



MFP Growth in the 1920s was 15 percent lower than the
benchmark 1929-41 period. During the 1920s it was
almost entirely a story about progress in manufacturing,
although advance was broadly experienced throughout the
sector, in contrast with the 1990s
In the 1930s, manufacturing’s contribution declined,
although remaining high in absolute terms. Transport and
Public Utilities played a much more important role. The
same was true to a lesser extent of distribution. These
latter effects reflect a roughly three decade long delay
between the invention of the internal combustion engine
and the full reaping of its productivity benefits in using
sectors.
MFP growth in 1995-2000 was three times what it had been
during the anemic years 1973-1995, but less than half the
rate clocked between 1929 and 1941.
Conclusions



In accounting for the expansion of potential
output, there is no single process, product, or
technological complex for the 1930s around
which one can build a compelling narrative.
What was exceptional about 1929-41,
distinguishing the period from the other two
examined in this paper, was the broad frontier
across which technological progress was
advancing.
In contrast, MFP advance 1995-2000 was
narrowly concentrated within a declining
manufacturing sector in SIC 35 and 36, and
within the using sectors in wholesale and retail
trade and securities trading.
A Century of Technological
Expositions

1893 –Chicago
• Columbian Exposition – The White City
• AC current for illumination, but still powered by coal and
steam

1939-40 – New York
• GM’s Futurama – Norman Bel Geddes’ wildly popular and
largely accurate vision of the US in 1960
• Democracity

1964-65 – New York
• Failure of the new Futurama – either to inspire or to
accurately forecast

1996 – Anaheim
• A Retro Tomorrowland
Is Comdex a Substitute?



The last “successful” North American world’s fair was Expo
’67 at Montreal, and the first (and last Asian fair Expo ’70
was also a success
Subsequent fairs (Spokane - 1974, Knoxville - 1982, New
Orleans - 1984, Vancouver - 1986, and Seville -1992) have
been more narrowly themed. Who now remembers them?
The fading impact of these international expositions
coincides roughly with the collapse of the residual
beginning in the 1970s.
World’s fairs emerged out of commercial fairs, and in the
1990s trade shows such as Comdex generated some of the
same kind of excitement as had earlier international
expositions, but they were much more narrowly focused,
reflecting the relatively narrow based of technical advance
in recent years, and have not been targeted at the general
public.
Table 9
Sectoral Contributions to Multifactor Productivity Growth Within the Private Nonfarm Economy
United States, 1929-1941
Sectoral
Manufacturing
Transport and Public Utilities
Wholesale and Retail Trade
Other Sectors
(net)
Mining
Construction
Finance, Insurance, Real Estate
(see note a)
Other Services (see note b)
TOTAL
1941 Share of
1941 Share of Private
National Income
Nonfarm Economy
MFP Growth
1929-1941
Sector's
Contribution to Aggregate
(PNE) MFP Growth
31.86
9.21
16.70
42.57
12.30
22.31
2.91
4.48
1.81
1.239
0.551
0.404
17.08
22.82
0.51
0.116
2.30
4.03
4.78
5.97
74.85
100.00
2.310