The Productivity Payoff of Computers:
A review of The Computer Revolution: An Economic Perspective by Daniel E. Sichel


Yannis Bakos

Stern School of Business, New York University

Science, July 3, 1998

In the last decade, American firms have made unprecedented investments in information technology, which, according to some estimates, now accounts for 30% of new capital investment. Anecdotal evidence for the benefits of information technology is easy to come by in our Internet age, as companies are transforming their operations, their marketing, and their relationships with customers and suppliers. Amidst the phenomenal IT revolution, several economists have been puzzled to find only modest growth in productivity reflected in the official statistics of the United States economy. The recognition of this "productivity paradox" is often attributed to Nobel-winning economist Robert Solow, who famously quipped in 1987 that computers can be seen "everywhere except in the productivity statistics."

To address issues raised by this paradox, researchers have studied the productivity of information technology at the level of individual firms. The coefficients of a "production function" are estimated from data about inputs (typically computer and non-computer capital stock, information systems labor, other labor, and R&D) and outputs (typically sales or value-added); this estimates the contributions of the included input factors. Brynjolfsson and Hitt (1995, 1996) analyzed data from 367 firms over the period 1988-1992 with 1,248 total data points, and found no evidence of productivity shortfall, and if anything, evidence of excess returns to IT capital and IS labor. Their results were robust to different specifications of their production function, and have been replicated by Lichtenberg (1995) and others with data from different sources.

Although these studies indicate that there is no "productivity paradox" at the firm level, macro-level data still show that measured economy-wide productivity gains have not substantially accelerated since 1960, despite rapidly increasing investments in computers and other types of information technology (e.g., see Roach 1992). Four prominent hypotheses have been proposed to explain this fact: mismeasurement, mismanagement, diffusion delay, and the capital stock theory. Mismeasurement suggests that a large proportion of the benefits from information technology will not show up in productivity statistics because they take the form of greater convenience, product variety, quality or timeliness—contributions that are largely missed in traditional GDP accounting (Baily and Gordon 1988, Griliches 1994). Mismanagement might lead to wasteful or unproductive information technology investments. The diffusion hypothesis suggests that many years may pass before the productive potential of an innovation is fully realized. Paul David (1990) offers the example of the electric dynamo, where the productivity benefits followed the required technical developments (largely complete by 1880) by several decades.

In his book, Daniel Sichel, a senior economist at the Federal Reserve Board, makes the case for the fourth explanation of the "productivity paradox," the capital stock theory he developed in collaboration with Stephen Oliner (Oliner and Sichel 1994). Providing a rigorous discussion of growth accounting for the U.S. economy that is accessible to the non-economist, Sichel describes how economists derive the output and contribution to economic growth for the hardware and software sectors. He then shows that despite the large current spending in information technology, computers are still only a small fraction of the existing capital stock. One reason for this is that only recently firms have been making substantial investments in computers, while they have accumulated large amounts of other productive capital from many decades of investment. Another reason is that computers rapidly become obsolete, and as a result large spending does not imply a large accumulation of capital stock.

In 1993, broadly defined "information processing equipment" was estimated to be 11.7 percent of the stock of nonresidential equipment and structures, and computers themselves accounted for a mere 2% of this capital stock (in nominal dollars). Sichel argues that since computers represent such a small fraction of capital stock, their contribution to economic growth should be correspondingly small. For example, he estimates that for the period 1980-1992, computer hardware contributed only 0.20% out of the total 2.3% average annual growth in gross nonfarm output. Sichel estimates the contribution of computers to economic growth through 2003 under different assumptions, projecting a contribution to net growth (i.e., after depreciation) ranging from 0.11 to 0.38% at the end of that period.

Sichel’s analysis is important, because we must set realistic expectations for the ability of IT to foster productivity and economic growth. It is equally important, however, to remember that even if Sichel’s estimates are accurate, the glass is not empty. With recent GDP growth in the U.S. around 2% and total productivity growth currently just above 1%, a 0.2% contribution of IT to economic growth would be nothing to sneeze at. The continuing improvement in the cost/performance ratio of IT is unprecedented, and as a result it is hard to point at any other single technology that has a comparable impact on productivity growth.

The real picture, however, is likely to be brighter than that painted in the book. For example, Sichel does not measure directly the contribution of IT to growth, but he infers it by assuming that the net return of computers is similar to that for other capital, around 12% annually. Econometric studies at the firm level suggest that marginal net returns to IT capital investments may be twice as high as this, and, since the ability of firms to deploy IT is limited by the availability of complementary factors such as IS labor, their average returns may be even higher.

Most important, Sichel’s approach may be too narrow. What is exciting about information technology is not its ability to substitute for other capital, but its ability to restructure every aspect of business, in the process creating new types of markets and organizations. Unfortunately Sichel’s analysis is not capable of capturing these higher-order impacts. Nevertheless, as long as his findings are kept in perspective, Sichel deserves credit for contributing an interesting argument and a rigorous methodology to our understanding of how information technology affects productivity and economic growth.




Baily, M. and R. Gordon, "The Productivity Slowdown, Measurement Issues, and the Explosion of Computer Power," Brookings Papers on Economic Activity, 2, 1988, pp. 347-431.

Brynjolfsson, E. and L. Hitt, "Information Technology as a Factor of Production", Economics of Innovation and New Technology, 3, 185-200, (1995).

Brynjolfsson, E. and L. Hitt, "Paradox Lost? Firm-level Evidence on the Returns to Information Systems Spending", Management Science, April, 42, (1996) p.541.

David, P. A. "The Dynamo and the Computer: An Historical Perspective on the Modern Productivity Paradox," The American Economic Review Papers and Proceedings, May 1990, pp. 355-61.

Griliches, Zvi. "Productivity, R&D, and the Data Constraint," The American Economic Review, March 1994, pp. 1-23.

Lichtenberg, F. R., "The Output Contributions of Computer Equipment and Personal: A Firm-Level Analysis", Economics of Innovation and New Technology, 3 201-217, (1995).

Oliner, S. D., and D. E. Sichel, "Computers and Output Growth Revisited: How Big Is the Puzzle?" Brookings Papers on Economic Activity, 2, 1994, pp. 273-334.

Roach, S. S. Technology Imperatives. Morgan Stanley, New York, 1992