Arnold Kling

Solow on the Solow Paradox

Arnold Kling, Great Questions of Economics
Previous Entry Next Entry

John S. Irons pointed out this interview with Nobel Laureate Robert Solow. There is material here for several posts. First, Solow continues to be skeptical about the role of computers in productivity.

from about 1995 until 2000 we had a period of much faster productivity growth. It is very tempting, it may even be plausible, but it's not a clear certainty, that some or all of that acceleration of productivity is the computer at last bearing fruit...

The best guesses—but they are merely best guesses—are that in approximately the next five to 10 years, we'll experience productivity growth nowhere near what we did toward the end of 1990s but maybe a half a percent to a percent a year faster than in the 1970s...Comparing the computer with electricity or the internal combustion engine just doesn't seem to me to be justified yet...

I always thought that the main difference the computer made in my office was that before the computer my secretary used to work for me, and afterward I worked for my secretary!

I am loathe to disagree with Professor Solow, but let me raise an alternative point of view. That is that in 1987, when Solow wrote his famous line that "we see the computer age everywhere but in the productivity statistics" we were still early in the life of Moore's Law, and the state-of-the-art Intel chip was the 386. To quote from one of my recent essays,

In fifteen years of progress from 1985 to 2000, the standard microprocessor went from the 386 with 275,000 transistors to the Pentium 4 with 42,000,000 transistors.

I find it plausible that somewhere along the way the computer went from being a tool that turned professionals into secretaries to being a tool that helped productivity. Thus, I would attribute the pattern of acceleration of productivity to the nonlinearity of Moore's Law rather than to luck. And what I see as the "best guess" about productivity going forward is Brad DeLong's optimistic oulook. But we'll know more in ten years.

Discussion Question. Think of the economy as having two sectors. The "normal" sector grows at 1.5 percent per year, and the computer sector grows at 20 percent per year. If the computer sector starts out at 0.5 percent of the economy, and the "normal" sector starts out at 99.5 percent of the economy, how long will it take for average economic growth to reach 2.0 percent per year?

Return to top