The Productivity Paradox: Rising Output, Stagnant Living Standards

Today's puzzler: Where did all the productivity go? Human work has become ever more efficient, thanks to the microelectronics revolution. My computer makes me a more productive writer. Factories now turn out better products with fewer workers. Even in the service sector, checkout clerks ring up twice the sales in half the time, using optical scanners.

But if the same labor can generate so much more output, why are living standards virtually stagnant? Why is a dramatic improvement in the productivity of capital not translating into faster economic growth? And why did the rate of productivity growth actually fall during the 1970s and 1980s?

There are five possible explanations, depending on how you frame the problem:

-- The lagging service sector. As Princeton University economist William Baumol has long noted, productivity growth in service jobs lags behind the growth in manufacturing productivity. Taxi drivers perform the same task today as they did in 1920, despite better cars. They may even do it more slowly, because of traffic jams. While advanced physical capital does make some service jobs more efficient, the waiters, nurses, and barbers simply do not become more productive over time, despite technological breakthroughs. So, as we become more of a service economy, our overall productivity growth will slow. Moreover, as other people's wages rise, reflecting improvements in the equipment they work with, service workers may actually become overpaid relative to their own static productivity, as their wages rise in tandem with living standards generally.

-- Unmeasured quality gains. An intriguing explanation is that we are actually living better than the statistics admit. Economist James K. Galbraith of the University of Texas at Austin, the father of two small children, observes that a teddy bear with a built-in cassette-player is neither more teddy bear nor better teddy bear. It is an entirely different product. But if it costs the same as an old-fashioned teddy, the consumer gets the benefit of an innovation. Similarly, if a compact disk costs $12.95, that is roughly the same in real dollars as the $2.95 that a long-playing record cost in 1953. So if we spend, say, the same 1% of our income on recorded music in 1993 as we did in 1953, we are unconsciously choosing to spend it on technically better audio rather than on other amenities--but we may not experience it as an explicit improvement in living standards or even as a conscious choice. The same is true of everything from bigger-screen, higher-resolution televisions and Ford Tauruses that don't break down to tastier, free-range chickens.

-- Income distribution. In the 1980s, more than half the growth in real income went to the top 1% of the population. This not only affects how the fruits of productivity gains are shared but also imposes some real economic costs. Economist David M. Gordon of the New School for Social Research contends that social disintegration has necessitated a dramatic increase in what he terms "guard labor"--workers in the public and private sector hired to compensate for social breakdown. The hiring of guard labor and the process of deterring, arresting, and incarcerating lawbreakers adds nothing to real wealth, but it soaks up money.

Moreover, if ordinary people have flat purchasing power, investment in new technology will be retarded for lack of customers. Investment follows purchasing power. Economist Leslie Nulty of the United Food & Commercial Workers International notes that despite optical scanners, the number of workers in supermarkets has not declined. Why? Because the supermarkets, seeing where the new purchasing power lies, have added gourmet and specialty departments--whose low-productivity workers fall into the Baumol category: They offer personal service to shoppers rich enough to buy expensive or exotic items, just as grocers did in 1893. Globalization only compounds the problem: As U.S. factory workers become more productive, they can't reap the gains, lest the jobs migrate to areas of cheaper labor.

-- Automation. In the early 1960s, economists worried that automation would create what was termed "technological unemployment." That crisis never materialized, however, because the 1960s were a decade of 4% annual growth, which kept generating jobs and customers for the new technology. But with the 2% growth of the 1990s, automation does cause net job displacement, and high unemployment waters down society's overall rate of productivity. So to maximize the conversion of productivity gains into living standards, we need high growth and full employment.

-- Delayed action. According to techno-optimists such as George Gilder, some latent technical improvements are about to bear fruit. U.S. leadership in software is only now starting to bring improvements in production. So the shakeouts and restructurings of the 1980s may mean dramatic gains in the 1990s.

Each of these views, I think, contains a piece of the truth. President Clinton's challenge is to help America translate the gains on the factory floor into broadly improved living standards.

Before it's here, it's on the Bloomberg Terminal.