May 28, 1993
"Zero-to-60 in 5 Nanoseconds"
San Jose Mercury News
By Timothy Taylor
<< Back to 1993 menu

IT MAY sound like heresy, but some skeptics doubt that computers add to economic productivity. Evidence for their nagging suspicions comes from all directions -- anecdotes, industry-level studies, and macroeconomic statistics. Anyone who has survived the ordeal of installing a new office computer system will understand the dangers here. Workers are required to take time away from their jobs to sit through dry-as-dirt training sessions. Records are carefully filed away in the computer -- and then no one knows how to access that information ever again.

Once the system is in place, middle managers are often bewitched by their new toys; they begin to produce and require more status reports and updates than ever before. Once upon a time, the paper companies were concerned that computers would usher in the "paper-less office." Actually, computers have led to far more paperwork.

Of course, a few horror stories about counterproductive technology don't make a definitive case. Anecdotes illustrate, but they don't prove anything. However, several academic studies have also found that computers don't contribute much to productivity.

In a recent "Weekly Letter" of the Federal Reserve Bank of San Francisco, senior economist Ronald Schmidt describes one study that looked at U.S. and European manufacturing firms, and found that investment in information technology had "little, if any, marginal impact on output or labor productivity while all other inputs into production including non-information technologies capital had significant impacts on output and labor productivity."

Another study, looking back over several decades, suggested that manufacturing "firms had over-invested in those information technologies, and that the costs of the equipment exceeded their benefits."

Other economic studies (not mentioned by Schmidt) have examined computers in service industries like banking, finance, insurance, and wholesale and retail trade, again without finding much connection between computerization and productivity. There seems to be some truth in the popular stereotype that the high-fangled computers in financial industries have mainly served to move money around in faster and trickier ways, often doing as much to destabilize and unsettle the economy than to support it.

The macroeconomic evidence signals yet another caution about computers and productivity; the 1980s were a time when business and individuals invested very heavily in computers, but productivity growth for the economy stayed stubbornly low. One wisecrack, generally attributed to Nobel laureate economist Robert Solow of MIT, is that: "We see computers everywhere, except in the productivity statistics."

But this case for the prosecution is far from unanswerable; the computer revolution has its share of defenders who advance two main arguments.

One response is to question the data on productivity. One basic measure of productivity is output per hour of labor; for purposes of comparison, "output" is usually measured by the total dollar value of what is produced.

Imagine that a new computer-aided design system allows companies to produce, say, higher quality cars -- perhaps with better use of interior space, greater safety, and better fuel efficiency -- but these improved cars sell for the same price as the old ones. According to the productivity statistics, the same amount of labor is producing the same dollar value of cars, so the computer added nothing. In short, productivity statistics often fail to capture higher quality.

This problem is especially severe in the service-oriented industries that have invested so heavily in computerization. Surely, it would seem that with automatic teller machines and access to accounts by telephone, productivity in banking must be higher than a decade ago? Logic says "yes," but the productivity statistics won't show it.

An alternate defense is to accept that computers haven't increased productivity much so far, but to argue that the economy takes decades to assimilate an invention as far-reaching as the computer. Paul David of Stanford University has pointed out an interesting historical parallel. The technology for the dynamo was available in 1900, but it literally took decades for factories to be electrified, so that they could take advantage of this new capability.

Similarly, Intel introduced the silicon microprocessor less than 25 years ago, in 1970. Even a decade ago, desktop computers were relatively rare, and those that did exist would today be considered as slow and obsolete as dinosaurs. On this perspective, the "information age" is still in its infancy. Our society is just beginning to redesign its manufacturing and services, its entertainment and finance, its patterns of commuting and living, to take advantage of information technology.

In the end, the case against computer-driven productivity is certainly not proven. In the end, it is difficult to believe that this remarkable technology won't raise the world's standard of living.

But the academic argument over computers and productivity has a real-world lesson. Too many people buy computers (and other new technologies) the way they buy cars: paying for zero-to-60 in five seconds, and the ability to drive 140 mph, and other features that they will probably never use.

In a world where the boundaries of technological possibility are continually expanding, it's worth remembering that sophisticated technology, in and of itself, is worth nothing. The interaction between technology and people is what counts.

<< Back to 1993 menu