Retraining won't suffice as technology advances: Computer Revolution May Be Devastating
By Robert J. Shiller, Mercury News, Posted on Sun, Mar. 21, 2004

Insecurity about jobs is a defining characteristic of our age. Two worries arise most often: globalization, which makes jobs migrate to poorer regions, and computer technology, which can make them disappear altogether.

These worries plague people of all incomes and ages, and in all countries. As shown by Mexico's laments about job losses to China, people in emerging countries worry as much as those in advanced countries. In response, politicians propose various job retraining or education programs, but rarely confront the real long-term issues.

Worries about globalization and the computer revolution ultimately boil down to the same thing, because globalization is mostly a consequence of new information technology. Computers and the Internet made it possible to send information almost without cost, and to do business across cultures and continents.

The worry is that information technology could now be on the bend of a "hockey stick" curve, where it will suddenly take off at much greater speed than before. The resulting changes will not all be welcome in economic terms.

Hockey-stick curves often occur in nature. Growth at a fixed percentage rate can look like a hockey stick. Five percent per year of $1 is only five cents, 5 percent of $10 is only 50 cents, and the curve appears flat for a while. But then the growth becomes massive: Five percent of a billion is $50 million.

Information technology is growing at a faster pace than 5 percent. In one interpretation of ``Moore's Law'' -- named after Intel co-founder Gordon Moore -- the amount of components on a computer chip, and therefore its performance, doubles about every two years. That is about a 40 percent annual growth rate. The economic impact of such an exponential growth path, starting from small beginnings and viewed over decades, will certainly look like one very large hockey stick.

Warning 50 years ago

Fears about computers eliminating jobs are, of course, not new. Fifty years ago, Norbert Wiener of MIT, a great 20th-century mathematician and pioneer of computer science, warned of the threat that computers posed to jobs.

The economic problems that Wiener worried about have not materialized so far. A half-century later, nothing really bad happened to any major segment of our population that can be blamed on computers. But that is because computers started from small beginnings, so that we have up to now been on only the flat part of the hockey stick. Now that a critical mass seems to have been reached, will we see a dramatically different effect on employment?

Of course, on average, advances in computer technology must be good news. Not only are computers productive, but they also serve people, rather than being their competitors. Powerful new computers will make the human race a whole better off. The problem is that real people are not just ``average,'' so the real question concerns how the benefits derived from computers are distributed. Also, it's true that the new technology can create new jobs -- though there's no guarantee that it will do this in sufficient force.

The core problem with information technology is that its economic benefits might be concentrated and that the wealth it creates may accrue predominantly to people who have subtle skills that computers can't duplicate, who adopt the technology early on or who have the right business connections.

Concentrated wealth

Older information technology has already created ``winner take all'' effects in some occupations. Phonograph records (an early example of information technology) created musical superstars who sell their services to millions of people, putting out of business countless local performers who found their talents obsolete. Television (another older form of information technology) did the same thing for actors and athletes.

But phonographs and television are limited compared with what is possible now -- or may be possible soon. In the new economic world, many more occupations may follow the route of discarded local musicians, resulting in wealth concentrations on a vastly greater scale than before.

The risks created by computer technology are real and frightening. They will not be seriously changed by retraining and re-education programs, which will not give most people the skills they need to remain more efficient than the new machines. Instead, these risks must be dealt with by fundamental changes -- changes that will not make individuals more productive than machines, but will allow society to manage risks better and to redistribute them.

For example, as the difference in rewards provided by the market to those with and without special skills grows, a progressive tax system to subsidize low-wage jobs becomes more necessary. Other likely steps include broadening the scope of private insurance policies, so that they cover some risks that generate inequality, such as the loss of the market value of human capital or of one's home.

More cushioning needed

Borrowing and lending institutions will need to change, so that they better cushion people against the risk of personal bankruptcy. Financial markets will have to be broadened, so that people and businesses can better hedge their risks.

It is important to start thinking about what changes lie ahead, because it may be harder to agree later on.

As long as people remain under a relative ``veil of ignorance,'' in the sense that they do not yet know how they will personally be affected, they may find it easier to agree that progressive taxes should be used to subsidize low-wage earners. When greater income inequality is a fait accompli, principled discussion may give way to naked class struggle between the new super-rich and those who have only their misery to lose.

Robert J. Shiller is professor of economics at Yale University, and is the author of Irrational Exuberance and, most recently, The New Financial Order: Risk in the 21st Century. He wrote this article for Project Syndicate.