US Productivity May Not Be What It Seems
FORTY-TWO years ago Zvi Griliches, an orphan from Lithuania, arrived in the United States carrying a cardboard suitcase filled with all his belongings. This week he gave the prestigious presidential address at the American Economic Association's annual meeting in Boston.
``For me to stand here today is to live the American dream,'' said the Harvard University economist.
Professor Griliches is regarded as one of the key developers of modern econometrics - the addition of sophisticated mathematics to economics. ``An elder statesman,'' said Robert Solow, the Nobel-prize-winning economist who introduced him.
In his talk, Griliches had something to say about the American economic dream - the view that continuous gains in productivity will result in each generation living better than the previous one.
Starting in 1968, statistics showed a slowdown in productivity growth in the US. After a mild upturn in the early 1970s, the world economy was hit by two successive oil-price shocks. These reduced economic growth rates in most of the industrial nations to levels well below those experienced in the 1960s and early 1970s.
The effect of those shocks wore off and the price of energy, after taking account of inflation, returned close to its earlier level. But, notes Griliches, productivity growth did not recover much.
So, as he recalls, economists, businessmen, and politicians began to wonder if something more basic was wrong with the economy. The number of patents granted in the US began to decline in the early 1970s, they noted. The proportion of total national output devoted to industrial research and development started to shrink in the mid-1960s. And companies were getting fewer patents per R&D dollar spent.
Was this the problem with productivity?
Griliches, a stickler with statistics, points out some ``discordant facts.'' Other countries, which did not experience declines in R&D, suffered even more severe drops in productivity. The slowdown in productivity in the US lingered in construction, finance, and other services where output measurement is notoriously difficult. And it turns out the drop in patent applications was a bureaucratic mirage. There was a budget crisis in the Patent Office and patents were not being dealt with as fast. Patent applications actually just leveled off, though R&D spending continued to grow in absolute numbers.
So does R&D not create as much productivity bang for the buck as it used to? Though the paper on which his address was based covers 62 pages with its extensive footnotes, graphs, and references, Griliches never really answers the question. ``We are caught up in a mixture of unmeasurement, mismeasurement, and unrealistic expectations,'' he says. ``The productivity situation is both better than we think and also worse.''
Likely there have been significant productivity advances in various service sectors of the economy that have not been measured, Griliches figures. Data collection is still based on an economy centered on agriculture and a growing manufacturing component. But more than half of the economy consists of service and government activities where measures of real output and relevant price factors are inadequate. Measuring a bushel of grain may be easy. How do you measure the output of a lawyer?
At the end of World War II, about half the economy was easily ``measurable'' by statisticians. By 1990, more than two-thirds of the economy had become ``unmeasurable,'' Griliches figures.
Moreover, R&D investment started to rise in the mid-1980s and patent applications started to grow again five years ago. This augurs well for the future, he says.
Business lobbyists, seeking tax breaks or other favors from Washington, do not usually talk about that shift.
The bad news is that statisticians may not have subtracted enough from the ``book values'' of the nation's stock of physical capital (buildings, machinery, etc.) and human capital (knowledge, skills, etc) to take account of more rapid obsolescence as a result of technological change and greater international competition. In other words, Griliches says, the nation is not as wealthy as it thought. But productivity growth, if based on this lower wealth number, would be higher than present measurements.