US Companies Double Spending On Computers - and It Pays Off

Recent study says companies are wise to invest in `information services,' since it adds more to productivity

AMERICAN business more than doubled its spending on computers and other information-processing equipment between 1960 and 1992. The share of ``information'' equipment in total producer investment in durable equipment rose from about 17 percent in 1960 to 36 percent by 1992.

Has this paid off? Most definitely, says Frank Lichtenberg, an economist at Columbia Business School in New York. ``Information services has helped keep [the United States] competitive,'' he says. ``We have to be careful and not get carried away by the notion that we are in the twilight. Our productivity gains are reasonably respectable.''

Professor Lichtenberg completed a study published recently by the National Bureau of Economic Research (NBER), a nonpartisan research group in Cambridge, Mass, that addresses the questions: What is the relative contribution of expenditures on information-services equipment and personnel to a company's output? Is it at least as high as those resources used in other ways by a firm to provide output?

The study found that information-services (IS) spending boosts a company's output (or reduces costs) about six times as much as the same amount spent on production workers. Computer experts are paid 2.5 to 3 times as much as production workers. But the resulting increase in output means ``they are more than worth it,'' Lichtenberg says.

Some earlier studies of the impact of computers on output and productivity have found the return on investment to be low and perhaps even negative. Others have found positive returns, but have not proved that IS expenditures do any better than spending on other production factors.

Lichtenberg, however, has employed the data obtained by two surveys of information-systems chief executives on the impact of companies' spending on IS on output during the period 1988-1991. One survey was by Informationweek magazine and the other by Computerworld.

These surveys enabled Lichtenberg to make comparisons between companies in the same industries with different levels of IS spending. The number of companies surveyed, all major users of IS, ranged from 190 to 458, varying between years.

Total investment in computer equipment and personnel amounts to about 10 percent of the total spent by the companies surveyed. However, computer capital and labor jointly contribute, or account for, about 21 percent of output, Lichtenberg calculates. So he concludes that IS expenditures are about twice as productive as ordinary capital and labor spending.

The number of computer specialists in industry ``has grown by leaps and bounds,'' Lichtenberg says. A study by the National Science Foundation found that they numbered 86,800 in 1976 and 439,700 a decade later. The average annual growth rate in the number of these jobs was 16.2 percent, compared with about 2 percent for total employment.

Though IS employment, even at the end of this period, accounted for only a small share of total employment, IS employment growth accounted for a larger contribution to output growth (not total output) in that decade than the growth in non-IS employment. One IS employee could be substituted on average for six non-IS employees without affecting output, Lichtenberg says.

Lichtenberg and some other economists hold that government measures of the productivity of the US economy do not adequately reflect the gains brought about by computers, especially in such service industries as banking and insurance.

Svi Griliches, a Harvard University economist, maintains that three-quarters of the investment in computers and other IS equipment has been made by sectors of the economy which are ``unmeasurable'' in terms of reckoning productivity gains. Economists have difficulty, for instance, in measuring the gains to productivity from automatic teller machines. These sectors include finance, construction, trade, and some other services.

``We may be just at the beginning of the computer era, early in the diffusion and learning stages, with most of the productivity contributions still to come, as we learn how to use computers more effectivity and integrate them more efficiently into the existing production structures,'' Dr. Griliches writes.