FEDERAL RESERVE BANK OF NEW YORK

I N E C O N O M I C S A N D F I N A N C E
June 2001 Volume 7 Number 6

Investing in Information Technology: Productivity Payoffs

for U.S. Industries
Kevin J. Stiroh

Although firms have invested billions of dollars in information technology to boost their
productivity, many analysts continue to question whether these investments do in fact lead
to productivity gains. An industry-level analysis of productivity performance provides robust
evidence of a link, showing that the industries experiencing the largest productivity acceleration
in the late 1990s were the producers and most intensive users of information technology.

The U.S. economy enjoyed a remarkable upsurge in the This edition of Current Issues moves beyond the
late 1990s as unemployment declined, inflation aggregate data to examine the recent productivity perfor-
remained in check, and perhaps most important, labor mance of the individual sectors and industries that make
productivity growth accelerated to rates not consistently up the U.S. economy. The article investigates how pro-
seen since the 1960s. Annual labor productivity growth ductivity growth in different industries has varied over
for the U.S. nonfarm business sector averaged 2.8 percent time and how the observed variation relates to IT capital
over the 1995-2000 period, double its average annual accumulation. Two empirical questions are at issue: First,
rate of growth for 1973-95 and just short of its 2.9 average are U.S. productivity gains confined to a few industries
for 1959-73. During the same period, U.S. firms made or shared by many? Second, are industry productivity
massive investments in information technology (IT), gains linked to the use of IT?
defined here to include computer hardware, computer Analysis of the industry-level data reveals that a
software, and telecommunications equipment. By year-end broad productivity resurgence took place after 1995,
1999, the value of the net stock of IT capital equipment with all principal sectors and a majority of industries
approached $900 billion. posting productivity gains. The analysis also shows that
The economy’s strong performance has spurred con- the industries experiencing the largest productivity
siderable interest in the link between the U.S. productivity acceleration in the late 1990s were the producers and
revival and the IT revolution. Almost all analysts agree most intensive users of IT—a finding that provides
that progress in the production of IT, exemplified by the direct evidence of information technology’s role in the
continuous decline of quality-adjusted IT prices, has con- U.S. productivity revival.
tributed directly to aggregate, or economy-wide, produc-
tivity gains. Some debate remains, however, about the Productivity, IT, and the Economy
productivity benefits from the use of IT. A large body of The questions addressed in this article have important
microeconomic studies and several recent papers using implications for the economy. Consider first the ques-
aggregate data conclude that IT investment and use— tion whether productivity gains have occurred in many
activities that economists refer to as “capital deepening”— industries or are concentrated in just a few. If productiv-
have contributed to labor productivity gains. Some skeptics ity increases have been widespread, then the productivity
argue, however, that the surge in aggregate productivity revival is likely to be more enduring. In contrast, if the
owes little to IT use and can instead be traced largely to IT increases have been concentrated in a single sector, then
production and cyclical factors.1 the revival may be vulnerable to a slowdown in that one
CURRENT ISSUES IN ECONOMICS AND FINANCE

sector. Moreover, if productivity increases have gains is buttressed by the fact that productivity is pro-
occurred in many industries, the resulting income and cyclical. That is, productivity tends to move with over-
economic gains would likely be distributed across all economic activity because of changes in resource
industries; if, however, productivity increases have been utilization, productivity shocks, increasing returns, or
more localized, the economic gains could be skewed reallocation effects. Consequently, part of the U.S. pro-
toward a few industries. ductivity resurgence likely reflects particularly strong
output growth during the late 1990s. Disagreement
Investigating the distribution of productivity gains is
exists, however, about how much of the recent produc-
also useful in evaluating the claim that the relatively
tivity surge reflects improvements in the underlying
narrow group of industries that produce IT equipment
trend and how much is attributable to cyclical forces.5
accounts for much of the aggregate improvement in
productivity. The strong productivity performance of
these industries is clear and can be seen both in the offi- Is the Productivity Revival Widespread?
cial productivity statistics and in the ability of these A useful way to assess the breadth of the U.S. produc-
firms to manufacture ever more powerful IT equipment tivity revival is to examine the productivity perfor-
at lower and lower prices. If these firms alone have mance of the sectors and industries that make up the
enjoyed productivity gains, however, one might con- U.S. private economy. Using 1987-99 data on real gross
clude that the IT revolution is somewhat disappointing. output and full-time-equivalent workers from the
Bureau of Economic Analysis (BEA), one can construct
This leads to the second question addressed in this a measure of labor productivity—real gross output per
analysis: Is there empirical evidence of a link between full-time-equivalent worker—for ten broad sectors and
IT use—rather than production—and productivity their sixty-one constituent industries. The breakdown
gains? Over the last few decades, firms have invested by sectors and industries follows the BEA classifica-
heavily in IT in the hope of improving profits and pro- tions, with the exception that manufacturing is decom-
ductivity. Potential gains from IT could be realized posed into a durable and a nondurable component.6
through a number of channels. Improved information
Casual examination of the aggregate productivity
flows within firms contribute to more efficient organi-
growth series suggests that a breakpoint occurred in
zations; better inventory management helps to prevent
1995, so productivity growth for the earlier period
factory downtime and increases product availability to
(1987-95) can be compared with productivity growth
consumers; low-priced IT systems—for example, auto-
for the later period (1995-99). 7 Chart 1 plots average
mated payrolls or account management systems—
annual productivity growth in the 1995-99 period
reduce the reliance on high-skilled labor.2 against that in 1987-95 for the ten broad sectors that
Several recent studies have concluded that both IT compose the private economy. Any sectors above the
production and IT use are driving the aggregate U.S.
productivity revival.3 If IT use really leads to productiv-
Chart 1
ity gains, one would expect to see a link between IT
Productivity Accelerated in Eight of Ten Broad Sectors
investment and productivity gains across industries.
Such a link would allow individual industries and the 1995-99 productivity growth (percent)
8
economy as a whole to produce more output and
Durable
implies a real economic benefit from the IT revolution. manufactuing
6 Nondurable
Some analysts have argued, however, that these gains manufacturing Wholesale
trade
are not likely to be large. In this view, information tech- Retail
4 trade
nology may primarily be used to reallocate market share FIRE

between competing firms (for example, when a tradi- Mining

2 Services
tional bookstore loses business to an on-line book- Transportation
and utilities
seller), replicate existing activities (when a retailer 0
offers both Internet and catalogue shopping), or Agriculture
increase on-the-job consumption (when workers play Construction
-2
video games or day-trade). Moreover, the substantial -2 0 2 4 6
training and support costs that often accompany IT 1987-95 productivity growth (percent)
investment may limit output gains. Indeed, if all of these
Sources: Bureau of Economic Analysis; author’s calculations.
forces are large enough, one might not see any link
Notes: All estimates represent average annual growth rates of real gross
between IT investment and productivity gains.4 output per full-time-equivalent worker. The diagonal line indicates no
change in productivity growth. Sectors above the line show productivity
Since actual productivity has in fact accelerated in growth acceleration; those below it show productivity growth deceleration.
recent years, the view that IT use has brought no real Sectors are weighted by their 1995 share of private employment.

FRBNY 2
diagonal line show an acceleration of productivity Chart 2
growth, while sectors below the line show a decelera- Productivity Accelerated in a Majority of Industries
tion of productivity growth. Since these sectors vary
1995-99 productivity growth (percent)
considerably in terms of size, the chart represents each 20 Security & commodity
with a plot point proportional to its 1995 share of pri- Holding & other brokers
vate employment. 15 investment offices Industrial
machinery Electronic & other
& equip. electric equip.
The chart shows a broad productivity revival across 10
virtually all of the private U.S. economy. Eight of the
ten major sectors experienced accelerating productivity 5
growth after 1995. 8 As one would expect, the durable
0
manufacturing sector, which produces IT hardware and
equipment, achieved especially impressive productivity
-5
gains after 1995, but many other sectors also showed
sizable gains. In particular, relatively large sectors such -10
as retail trade, services, nondurable manufacturing, and
finance, insurance, and real estate (FIRE) showed faster -15
-15 -10 -5 0 5 10 15
productivity growth in the late 1990s. The two sectors
1987-95 productivity growth (percent)
that experienced a deceleration, agriculture and mining,
are relatively small, accounting for just 2.9 percent of Sources: Bureau of Economic Analysis; author’s calculations.
private output in 1999. Notes: All estimates represent average annual growth rates of real gross
output per full-time-equivalent worker. The diagonal line indicates no
The change in the average productivity growth rates change in productivity growth. Industries above the line show productivity
from the earlier to the later period differs considerably growth acceleration; those below it show productivity growth deceleration.
across sectors, ranging from -1.25 percentage points in
agriculture to 2.50 percentage points in durable goods it is quite difficult to differentiate trend from cycle
manufacturing. This finding suggests that looking only without observing a full business cycle. Second, each
at aggregate data may obscure important differences industry is likely to have different cyclical properties,
within the economy. so any attempt to control for these effects could intro-
The data on productivity growth in individual indus- duce considerable noise from an imperfect adjustment
tries yield findings similar to the sectoral productivity procedure. Third, the recent productivity revival
results. Chart 2 plots average annual productivity appears somewhat different from its predecessors, so it
growth rates for the sixty-one industries in the BEA may be inappropriate to apply adjustments based on his-
classification for the same periods as in Chart 1. Again, torical relationships to the current period. For example,
productivity gains appear widespread, with thirty-eight most of the postwar productivity revivals have occurred
of the sixty-one industries showing faster productivity as the economy exited recession, while the current
growth in the 1995-99 period than in the 1987-95 period of rising productivity growth began very deep
period. For the sixty-one industries, the mean increase into the economic expansion.11
in productivity growth was 1.09 percentage points; the Overall, the results presented thus far point to a
median acceleration was 0.60.9 broad productivity revival in the late 1990s that encom-
Two IT-producing industries—industrial machinery passed most industries and sectors in the private econ-
and equipment, which includes the manufacture of com- omy. While the analysis does not establish whether the
puter hardware, and electronic and other electric equip- productivity gains should be attributed to cyclical
ment, which includes the production of semiconductors forces or to the changes in the underlying trend, it
and telecommunications equipment—showed exceptional demonstrates that the recent productivity revival is not
limited to a few industries that produce IT or other
gains. These gains are largely attributable to the rapid
durable goods.
technological advances that are driving the IT revolution.
In addition, two finance-related industries—security and
commodity brokers and holding and other investment Is the Productivity Revival Linked to IT?
offices—experienced a sharp acceleration in productivity The second question addressed in this analysis is
growth. The large gains recorded for these industries may whether the broad productivity revival is linked to the
be an artifact of how the BEA measures output, and massive investment in IT. From 1996 to 2000, U.S.
firms spent nearly $2 trillion on IT hardware and soft-
therefore productivity, in these industries.10
ware in pursuit of increased efficiency, higher produc-
Note that, for several reasons, the data in Charts 1 tivity, and stronger profits. The potential benefits from
and 2 have not been adjusted for cyclical effects. First, IT investment vary enormously, however, across sectors

3
 CURRENT ISSUES IN ECONOMICS AND FINANCE

and industries. For example, IT may be a very valuable One can take advantage of this wide variation in IT
tool in a financial firm, but it is likely to have fewer intensity to assess the impact of IT use on productivity.
applications on a farm. Thus, a natural first step in If IT accumulation does indeed contribute to productiv-
seeking a link between productivity gains and IT is to ity, one would expect the most intensive users to show
assess the intensity of IT use in different parts of the the largest productivity gains. In contrast, if the U.S.
economy. productivity revival is not an IT phenomenon, industry-
level productivity gains would likely be independent of
Chart 3 plots IT intensity, measured as the current
IT accumulation.
dollar share of IT capital in total reproducible, nonresi-
dential assets, for each of the ten major sectors in 1999.12 Before the analysis proceeds, however, two important
As with the sectoral productivity data, considerable varia- issues must be considered. First, how should IT intensity
tion is evident in IT intensity across sectors. In wholesale be defined? A useful definition is the one introduced
trade, 20.7 percent of the value of the capital stock was in above, namely, the share of IT capital stock in the repro-
IT assets in 1999, while agriculture had only a 0.9 percent ducible capital stock, because it captures the investment
share. Other IT-intensive sectors were the transportation resources allocated toward these high-tech assets. Second,
and public utilities sector, which includes communications how can one control for potential reverse causality?
industries (13.0 percent); services (11.7 percent); FIRE Although a link between IT and productivity growth
(7.8 percent); and durable manufacturing (7.6 percent). could reflect the fact that IT contributes to productivity,
the causality could also run the other way, because
Even more pronounced differences in IT use
industries with strong productivity growth might make
emerged across industries. Chart 4 plots the distribution
large investments in IT. One way to resolve this problem
of IT capital shares in 1999.13 The average capital share
is to compare IT intensity with subsequent productivity
was 8.9 percent and the median was 4.8 percent, with a
growth—for example, by comparing IT intensity in
range from 0.4 percent (farms) to 38.6 percent (tele-
1995 with productivity growth in 1995-99. While this is
phone and telegraph). Other industries with a signifi-
not a perfect control if firms have serially correlated
cant share of their capital stock in IT assets include
productivity shocks or make investment decisions today
radio and television (34.0 percent), transportation ser-
in anticipation of future productivity shocks, it does
vices (30.1 percent), business services (29.7 percent),
lessen the endogeneity concern.14
nondepository institutions (27.8 percent), legal services
(23.5 percent), motion pictures (22.9 percent), and To test for a link between productivity and IT pro-
wholesale trade (20.7 percent). These industries with IT duction and use, one can compare the change in average
capital shares above 20 percent attest to the rapid accu- productivity growth from 1987-95 to 1995-99 for three
mulation of IT in industries where this technology has sets of industries: industries that produce IT, industries
many useful applications. that use IT intensively, and other industries (see table).
As mentioned earlier, fundamental technological
progress has enabled the IT-producing industries—
Chart 3 industrial machinery and equipment and electronic and
Sectors Varied Markedly in Intensity of IT Use in 1999
IT capital stock share (percent)
25
Chart 4
Industry IT Capital Shares Ranged Widely in 1999
20 Percentage of industries
40
Private economy
15 average = 9.1%
30
10

20
5

0
. . 10
e ng on fg fg l. ale il s
ltur ini ti m m uti s ta RE ice
ricu M tr uc ur. ur. s.
& hole Re FI
Serv
Ag ns D nd an W
Co No Tr 0
0 5 10 15 20 25 30 35 40
Sources: Herman (2000); author’s calculations. IT share of reproducible, nonresidential capital (percent)
Note: The IT capital stock share is defined as the 1999 current-dollar IT
capital stock as a percentage of the reproducible, nonresidential capital stock. Sources: Herman (2000); author’s calculations.

4 FRBNY
other electric equipment—to achieve very strong pro- goods, from numerically controlled machine tools to coffee
ductivity gains, so it makes sense to isolate these indus- makers. In addition, the data on IT use do not specifically
tries. IT-intensive industries are then defined as other account for complementary innovations such as the orga-
industries with an IT capital stock share above the 1995 nizational restructuring or skill changes that often accom-
median of 3.8 percent; the third category consists of all pany IT investment. Thus, to the extent that other factors
remaining industries. are correlated with IT investment, the IT capital share
should be thought of as an indicator for all of the changes
The two IT-producing industries show a mean pro-
that accompany the IT revolution rather than the precise
ductivity acceleration of 3.7 percentage points, while
impact of IT capital alone.
the twenty-six IT-intensive industries show a mean pro-
ductivity acceleration of 2.0 percentage points. In sharp
contrast, productivity gains for the other twenty-nine Conclusion
industries averaged only 0.4 percentage point. When the The sharp acceleration of U.S. labor productivity
industries are weighted by their relative size to better growth and steady accumulation of computing and
represent the economic impact of each industry, the communication power have led many to believe that IT
results are similar.15 The IT-producing industries show a is a driving force behind the U.S. productivity revival.
gain of 3.7 percentage points, the IT-intensive industries By underscoring the wide variation in both IT intensity
a gain of 1.4 percentage points, and the other industries and productivity growth across U.S. industries and by
a gain of only 0.1 percentage point. These results suggest showing a link between the two, this industry-level
an important link between IT use and productivity gains. analysis supports that view.

Two caveats, however, are in order. First, one can argue The analysis also suggests that the U.S. productivity
that cyclical factors are imperfectly controlled for in this revival is a real phenomenon, not just a cyclical one.
type of analysis. Since some of the recent productivity Given the large differences in productivity gains
acceleration is undoubtedly a result of strong output between IT-intensive and other industries in the late
growth, there is some merit to this claim. Given the wide 1990s, cyclical forces would have to be highly concen-
variation in productivity that appears to be linked with trated in precisely those industries that are most IT-
lagged IT intensity, however, IT use likely plays an impor- intensive to qualify as the whole story. While informa-
tant role. Second, IT intensity in this analysis is defined tion technology cannot explain everything about the
by investment in computer hardware, computer software, U.S. productivity revival, the robust link between IT
and telecommunications equipment. Yet there is much intensity and productivity gains suggests that there is an
more to the IT revolution. Semiconductors, for example, important economic relationship.
are now routinely embedded in many other types of
Notes
IT Producers and Users Show Largest 1. The micro studies are surveyed by Brynjolfsson and Hitt (2000).
Productivity Gains Aggregate studies that report an impact from IT use include BLS
Average Annual (2000), CEA (2001), Jorgenson (2001), Jorgenson and Stiroh
Productivity Growth (2000), Oliner and Sichel (2000), and Whelan (2000). IT skeptics
(Percent) include Gordon (1999, 2000) and Kiley (1999, 2000).
Number of
Industries 1987-95 1995-99 Change 2. See Brynjolfsson and Hitt (2000) for several case study examples
Unweighted averages of IT benefits.
IT-producing industries 2 8.53 12.22 3.69 3. See BLS (2000), CEA (2001), Jorgenson (2001), Jorgenson and
IT-intensive industries 26 1.18 3.16 1.99 Stiroh (2000), and Oliner and Sichel (2000).
Other industries 29 1.87 2.30 0.43
4. See Baily and Gordon (1988), Kiley (1999, 2000), and Gordon
Weighted averages (1999, 2000) for details.
IT-producing industries 2 8.24 11.90 3.66
5. See CEA (2001) and Gordon (2000) for alternative estimates.
IT-intensive industries 26 1.24 2.61 1.37
Other industries 29 0.98 1.11 0.13 6. The data are all from the BEA’s gross product originating data-
base, described by Lum and Moyer (2000). The industries are at
Source: Author’s calculations. roughly the two-digit level as defined by the Standard Industrial
Notes: Unweighted values for each group of industries are calculated as the Classification (SIC) system.
mean of the industries’ annual productivity growth rates for each period. Weighted
estimates use the share of private employment to weight the productivity growth 7. More formal econometric work in Stiroh (2001) points to a
rates of each industry. The IT-producing industries are industrial machinery and break in the aggregate productivity series in third-quarter 1995.
equipment and electronic and other electric equipment. IT-intensive industries
have a 1995 IT capital share above the 1995 median. Four industries are excluded 8. CEA (2001), Nordhaus (2000), and Stiroh (2001) also find a
because detailed capital stock data are not available. broad productivity revival using value-added data.

5 FRBNY
CURRENT ISSUES IN ECONOMICS AND FINANCE

9. Econometric tests in Stiroh (2001) show that these productivity Gordon, Robert J. 1999. “Has the ‘New Economy’ Rendered the
gains are statistically significant. Productivity Slowdown Obsolete?” Unpublished paper,
Northwestern University, June 12.
10. See Lum, Moyer, and Yuskavage (2000) for details on how out-
put is measured in these industries. ———. 2000. “Does the ‘New Economy’ Measure Up to the Great
Inventions of the Past?” Journal of Economic Perspectives 14,
11. Basu, Fernald, and Shapiro (2000) conclude that the recent pro-
no. 4 (fall): 49-74.
ductivity acceleration stems from faster technological change, not
from transitory factors like factor utilization and factor accumulation. Herman, Shelby W. 2000. “Fixed Assets and Consumer Durable
Goods for 1925-99.” In Bureau of Economic Analysis, Survey of
12. These data are based on the BEA’s Tangible Wealth Survey,
Current Business, September: 19-30.
reported in Herman (2000).
Jorgenson, Dale W. 2001. “Information Technology and the U.S.
13. The distribution in Chart 4 shows the fraction of industries with
Economy.” American Economic Review 91, no. 1 (March): 1-32.
IT capital shares in a given range. Note that detailed capital data are
not available for four industries—social services, membership orga- Jorgenson, Dale W., and Kevin J. Stiroh. 2000. “Raising the Speed
nizations, other services, and private households—so the analysis of Limit: U.S. Economic Growth in the Information Age.”
industry IT is limited to fifty-seven industries. Brookings Papers on Economic Activity, no. 1: 125-211.

14. One could also implement other timing conventions or instru- Kiley, Michael T. 1999. “Computers and Growth with Costs of
mental variable techniques, as in Stiroh (2001). Adjustment: Will the Future Look Like the Past?” Board of
Governors of the Federal Reserve System, Finance and
15. These estimates are taken from a weighted least squares regres-
Economics Discussion Series, no. 1996-36, July.
sion with dummy variables for the different means and accelerations
across the three types of industries, with employment as weights. ———. 2000. “Computers and Growth with Frictions: Aggregate
and Disaggregate Evidence.” Unpublished paper, Board of
Governors of the Federal Reserve System, October.
References
Lum, Sherlene K. S., and Brian C. Moyer. 2000. “Gross Domestic
Baily, Martin Neil, and Robert J. Gordon. 1988. “The Productivity Product by Industry for 1997-99.” In Bureau of Economic
Slowdown, Measurement Issues, and the Explosion of Computer Analysis, Survey of Current Business, December: 24-35.
Power.” Brookings Papers on Economic Activity, no. 2: 347-420.
Lum, Sherlene K. S., Brian C. Moyer, and Robert E. Yuskavage.
Basu, Susanto, John G. Fernald, and Matthew D. Shapiro. 2000. 2000. “Improved Estimates of Gross Product by Industry for
“Productivity Growth in the 1990s: Technology, Utilization, or 1947-98.” In Bureau of Economic Analysis, Survey of Current
Adjustment?” Paper presented at the Carnegie-Rochester Business, June: 24-54.
Conference on Public Policy, Pittsburgh, Pa., November 17, 2000.
Nordhaus, William D. 2000. “Productivity Growth and the New
Brynjolfsson, Erik, and Lorin Hitt. 2000. “Beyond Computation: Economy.” Unpublished paper, Yale University, November 20.
Information Technology, Organizational Transformation, and
Oliner, Stephen D., and Daniel E. Sichel. 2000. “The Resurgence of
Business Practices.” Journal of Economic Perspectives 14, no. 4
Growth in the Late 1990s: Is Information Technology the
(fall): 23-48.
Story?” Journal of Economic Perspectives 14, no. 4 (fall): 3-22.
Bureau of Labor Statistics (BLS). 2000. “Multifactor Productivity
Stiroh, Kevin J. 2001. “Information Technology and the U.S.
Trends, 1998.” USDL 00-267, September 21.
Productivity Revival: What Do the Industry Data Say?” Federal
Council of Economic Advisers (CEA). 2001. “Annual Report of the Reserve Bank of New York Staff Reports, no. 115, January.
Council of Economic Advisers.” In Economic Report of the
Whelan, Karl. 2000. “Computers, Obsolescence, and Productivity.”
President. Washington, D.C.: U.S. Government Printing Office,
Board of Governors of the Federal Reserve System, Finance and
January.
Economics Discussion Series, no. 2000-06, January.

About the Author

Kevin J. Stiroh is a senior economist in the Banking Studies Function of the Research and Market
Analysis Group.

The views expressed in this article are those of the author and do not necessarily reflect the position
of the Federal Reserve Bank of New York or the Federal Reserve System.

Current Issues in Economics and Finance is published by the Research and Market Analysis Group of the Federal
Reserve Bank of New York. Dorothy Meadow Sobol is the editor.