Quick Links
Authors
While we live in the Information Age, tangible items like bricks and mortar, stores and offices, factories and mines, drill presses and lathes cannot be ignored. Physical capital remains vital to our nation's (or any nationss) GDP, productivity, wages and employment. The policy emphasis on more capital spending, in the U.S. and elsewhere, is not misplaced. Nonetheless, in a real sense, it is ideas that rule the world. So the U.S. and other governments interested in promoting growth should pay at least equal attention to policies that foster technology and innovation.
Perhaps surprisingly, this has always been so - or at least for as long as we have data. From the earliest studies of what is called "growth accounting" by Robert Solow and Edward Denison in the 1950s, right up to the latest studies, economists have consistently found that what is called "the residual" contributes more to economic growth than either growth of labor input or growth of the capital stock - sometimes more than both of them combined. Because "the residual" is the part of growth that cannotbe accounted for by increasing inputs of labor and capital, it is not a stretch to identify it with technology, which is exactly what generations of economists have done.
As Robert Shapiro and Nam Pham remind us in this fascinating study, an increasing share of the market valuation of the top U.S. companies is now apparently based on "intangibles" ("ideas," if you will), rather than on companies' stockpiles of physical assets. They reckon that this share of value rose from about 25 percent in 1984 to about 64 percent in 2005 - a huge increase in just two decades. It is true that market valuations of ideas can sometimes run amok, as the late 1990s illustrated dramatically. But, irrational exuberance aside, there is genuine, lasting value in ideas - at least for society. For example, while almost all of the dot.coms of 1999 have disappeared into the sands of history, the economic value of the Internet is huge and growing.
One reason why ideas are slippery things to value is that they are slippery things to measure. Mere counting certainly will not do, even if we had some way to count ideas - for, plainly, there is an immense difference in value between a good idea and a useless one. So economists look for proxy measures, of which expenditures on research and development (R&D) is probably the leading candidate. Although the process is inherently risky and random, more intense searching for ideas (as measured by R&D spending) will probably turn up more good ones. And an impressive amount of economic research, dating back decades, supports this hypothesis. Shapiro and Pham follow in this rich tradition. This body of research, by the way, also shows that R&D expenditures have, on average, offered high rates of return.
While the federal government spends enormous sums on R&D, that impressive expenditure is dwarfed by private sector R&D spending - on which Shapiro and Pham concentrate, focusing on the manufacturing sector. Which industries do the spending? All of them - but far from equally. Shapiro and Pham neatly divide major U.S. manufacturing industries into high- R&D and low-R&D sectors according to how much they spend on R&D per employee. The division is surprisingly clean and sharp. If we ignore the ever-present (but uninterpretable) "miscellaneous" category, the highest spender among the low-R&D industries (electrical equipment) spent an average of about $5,600 per employee on R&D over the years 2000- 2004. By contrast, the lowest spender among the high-R&D industries (basic chemicals) spent about twice as much. According to Shapiro and Pham, the five biggest-spending industries, in order, were (with spending per employee in parentheses):
Note that the rankings drop off significantly after second place.
Shapiro and Pham show that the high-R&D industries stand out in other respects, too. They had higher value added per worker; they paid higher wages; and they lost jobs at a slower pace. (The manufacturing sector as a whole lost 16 percent of its employment over the 2000-2004 period; among its specific industries, jobs increased only in pharmaceuticals.)
The sharpest correlation of all is with value added per employee, which is a rough measure of productivity. To a remarkable degree, America's most productive manufacturing industries are the ones that invest the most in R&D. How strong is that relationship? The picture on the next page should be worth a thousand words. It plots R&D per employee horizontally and value added per employee vertically; and the positive correlation is visible to the naked eye. Two industries stand out from the others with extremely high value added per employee: petroleum and pharmaceuticals. The story with petroleum is simple: The high price of oil enabled the industry to produce huge value added with little R&D. If we remove this one outlier, the "correlation coefficient" rises to a high value of +0.79; and the relationship looks roughly linear.
R&D Per Employee and Value-Added Per Employee Relationship Plotted
Shapiro and Pham conclude that U.S. policymakers should foster the creation of more intellectual property (IP) and work harder to protect the IP that American companies already have. However, with or without IP protection, good ideas will be copied, modified, improved upon and invented around. Today's fresh new invention is destined to become tomorrow's stale old idea. If America is to remain the leader of the economic pack, we must keep on innovating in the future - just as we have done in the past. Shapiro and Pham are right that the incentive to create IP depends, in part, on our ability to protect it. But it also depends on other things, such as entrepreneurship, a sensible tax system, a steady flow of scientific and engineering talent, vibrant capital markets and government support for basic science. Together, these and other ingredients comprise the raw materials for faster economic growth, higher productivity and higher wages.
Alan S. Blinder
Princeton University
May 2007