De Masi: During the 1990s, strong economic growth in the United States, combined with low inflation and a pickup in labor productivity growth, led many people to label the phenomenon a “new economy.” But there’s really little consensus on what is different about the U.S. economy or on what the term means.
Discussions have broadly raised three questions: Have there been fundamental changes in the trade-off between inflation and unemployment? Has the pickup in total factor productivity growth been spurred by spillovers and networking between various businesses? And, more narrowly, has technology triggered increased labor productivity growth? People have been puzzled by how U.S. unemployment declined so sharply without triggering inflation. For economists, this suggests that the nonaccelerating inflation rate of unemployment (NAIRU) may have declined, allowing continued strong growth without inflationary pressures. But at present it is very difficult to say whether the outstanding U.S. unemployment performance is permanent or temporary. Another question is whether one firm’s investment in information technology (IT) helps other firms become more productive. But networking effects are extremely difficult to measure, and to date there is little solid evidence to suggest that feedback effects are more prevalent across industries now than they were before. Most empirical research has focused on whether the observed acceleration in labor productivity growth is permanent and whether it can be linked primarily to IT.
De Masi: Recent studies have linked the pickup in U.S. labor productivity growth to the dramatically expanded production and use of IT that has occurred since the mid-1990s. Investment in computers and peripherals has surged by an average annual rate of 45 percent—partially driven by sharp declines in computer prices, which fell by about 24 percent a year. The IT sector, which accounts for a relatively small share of overall U.S. GDP, has seen its contribution to output growth double from the first to the second half of the 1990s, rising to a little over 1 percentage point. And U.S. labor productivity growth has surged—picking up by about 1 percentage point to about 2½ percent a year.
To see the key role of IT in all this, labor productivity growth can be separated into capital deepening—the amount of capital used per hours worked—and total factor productivity growth—that is, such things as improvements in production processes. In the second half of the 1990s, greater efficiencies in producing computers—particularly semiconductors—boosted total factor productivity growth. And sharply lower computer prices encouraged firms across the economy to raise their investment in technology, which contributed to capital deepening and further boosted labor productivity growth. Recent estimates suggest IT accounts for about two-thirds of the observed pickup in U.S. labor productivity growth in the latter half of the 1990s. The bottom line is that the impact of this technology may be a new development, but the process by which it has affected labor productivity growth is the same “old economy” process we have observed whenever new technologies have been introduced into the economy.
De Masi: That’s a crucial question right now. It’s highly uncertain whether strong productivity growth will continue, but there are promising indications that a portion of this accelerated growth rate in productivity will remain, at least in the near future. During a downturn, of course, a deceleration in productivity growth can be expected for cyclical reasons. But since technology appears to be driving the pickup in labor productivity growth and, given the strong rate of investment in technology over the past few years, a sudden drop-off in labor productivity growth is not expected.
Cerisola: The Canadian economy is highly, and increasingly, integrated with the U.S. economy, but labor productivity has not accelerated in Canada and, given the degree of integration, this is puzzling. Canada has not experienced the capital deepening that the United States has, nor has it seen an acceleration in trend total factor productivity growth. Cyclical, as well as structural, factors help explain this. The 1990–91 recession was much more severe and prolonged in Canada than in the United States; recovery was significantly slower; and investment in machinery and equipment was affected. Since 1996, however, investment has been as buoyant in Canada as in the United States, but we have not yet seen the upturn in capital deepening that the United States has had. Growth in the past two years has been much more labor-intensive in Canada.
The Canadian economy also faced several important structural changes—free trade agreements and significant deregulation in its telecommunications, transportation, and financial sectors—that have led to corporate restructuring. This restructuring may not only have delayed the recovery, it may also have contributed to the slow growth in total factor productivity, possibly because of significant lags between the restructuring and the gains in production efficiencies associated with it. The gap in labor productivity growth can largely be traced to developments in the manufacturing sector, where the relative importance of the IT sector is significantly smaller in Canada than in the United States. We have seen significant adoption of IT in Canada over the past few years, and although Canada is not as important a producer of IT goods as the United States, productivity growth has been strong in many other Canadian manufacturing industries and sectors.
Kodres: The story there is similar to Canada’s. For most of the euro area, labor productivity growth dropped in the latter 1990s. The region has experienced relatively high growth with low inflation in recent years, but it has not yet seen an acceleration in labor productivity growth. Some of this lag may simply be measurement error, however. When data discrepancies are taken into account, much, but not all, of the difference between the United States and Europe disappears.
An important issue is quality-adjusted prices. The United States adjusts for quality in the prices of computers and software, but only a few European countries do. Without quality-adjusted prices, Europe’s real investment in IT equipment appears much lower. For example, when the Bundesbank used U.S. price deflators to recalculate real annual expenditures on information technology since 1991, it found German investment increased to some 27° percent annually versus the 6 percent recorded using its national statistical methodology. When we attempted to adjust for measurement difficulties, we found that the contribution of IT investment to European output was closer to that in the United States and not as low as we had originally thought. European growth may be a bit higher too. But all this said, there are reasons to believe that Europe, even more than Canada, is behind the curve in the new economy, and this may be partly due to smaller sectors for computer production.
Estevao: Even France, which uses a measurement methodology similar to that in the United States, isn’t seeing much of a contribution to labor productivity growth from IT. This is Robert Solow’s U.S. productivity paradox of the 1980s: computers were everywhere but in the productivity statistics. The answer, of course, was that computers were not everywhere. IT was too small a proportion of total U.S. capital stock in the 1980s to register a presence. Europe is in that situation now. France and Germany have been investing in IT capital, but they got a late start. Give them five years. If they continue to invest in this way, you will see bigger contributions to productivity growth from capital deepening. We are not going to see much of a contribution from increased total factor productivity growth in the IT sector, however, because, as in Canada, this sector is a very small portion of total production.
Kodres: Europe’s lower rates of capital deepening are also associated with very large drops in unemployment and increases in employment. In Europe, much of this increased employment has come at relatively low skill levels. When you integrate lower-skilled labor into the existing capital stock, labor productivity is likely to decline. The mystery is the drop in U.K. labor productivity, where the bias in the composition of employment is not present. But in the United Kingdom, like Canada, the manufacturing sector is the key. Its productivity has been particularly low, with little capital deepening or investment. But the United Kingdom is now seeing a very sharp pickup in productivity in the service sector and in select areas in manufacturing potentially linked to technology.
Estevao: If you look at France, a lot of the deceleration in labor productivity growth is due to a deceleration in capital deepening. Faced with a large pool of unemployed, France has, since 1993, taken steps to lower labor costs through targeted reductions in labor taxation. Firms have responded by changing their optimal allocation between capital and labor.
Kodres: The United Kingdom also has very low unemployment. We are likely to see increases in labor productivity growth there, because firms will begin to substitute capital for labor as labor becomes more expensive. Germany never experienced the big drop in productivity that France and the United Kingdom did. German productivity gains have traditionally been at the high end of the scale for Europe, and the country has done an excellent job integrating technology into its manufacturing processes. Germany is known for taking existing theory and putting it into practice and making it work. And they’ve made it work via the old economy means of integrating new technology into large manufacturing companies.
Estevao: What’s striking is that all of these stories are very much old economy stories. The type of capital is different, but the story is very similar. We don’t see much hard evidence of changes in business behavior in major European countries. For instance, in contrast to the U.S. experience, total factor productivity growth has not yet picked up. The accompanying total factor productivity gains and lower inflation that are expected to result from Internet e-commerce and business-to-business (B2B) synergies have not really shown up in the data yet, though there is a lot of anecdotal evidence that suggests these synergies are present. So as of now, the little increase in the contribution of IT to output growth in France, Germany, and the United Kingdom has been due to an increase in IT capital deepening. As I said, it’s an “old story.”
Kodres: Germany, and most European countries, could find it useful to reduce the red tape small businesses face when they start up and to allow greater labor market flexibility. They might also consider giving greater encouragement to venture capital and to moving beyond bank-related financing opportunities.
Cerisola: The literature stresses that new investment and the speed with which it spreads in the economy tend to be linked with flexible labor markets and other issues, such as tax distortions. In Canada, labor markets tend to be very flexible by as suggested the Organization for Economic Cooperation and Development, though aspects of its employment insurance system may have constrained the functioning of its labor market and increased relative labor costs. The tax burden on capital and labor has been very high by international standards, especially relative to the United States, and this may have contributed to lower returns on investment. Recently announced reductions in taxes on personal and corporate income and capital, however, will tend to lower tax rates and eliminate some distortions. These steps are likely to encourage even greater investment in, and adoption of, technology.
Kodres: It may reflect the size of the country and the way enterprises are organized. A lot of small start-ups fail, but a good chunk of them are bought out. The ability to buy out a company or dissolve it is less common in continental Europe or even the United Kingdom.
Estevao: Probably not to the same degree, at least as long as Europe is not a major producer of computers and hardware. And the timing of an acceleration will depend on when the integration of lower-skilled workers comes to a natural halt. Policymakers have been focusing, rightly, on incentives to hire. And the tax burden is still quite high in many European countries. In the U.S. economy, people are paid a lot for a lot of work. But in France, this is taxed heavily. So we are not able to close this story yet. We see high rates of investment in France, but impediments to start-ups and relatively rigid labor markets.
Kodres: There appear to be lower levels of entrepreneurship in Germany and the United Kingdom, compared to the United States, but I am optimistic, given their histories of adopting new technologies and integrating them into manufacturing and the service industry. It will be a long, slow process, and we will have trouble seeing it in the data, but over the next five years, we will probably see a significant introduction of IT in Europe and a pickup in labor productivity growth.
Ian S. McDonald
Senior Editorial Assistant
Art Editor/Graphic Artist
The IMF Survey (ISSN 0047-083X) is published in English, French, and Spanish by the IMF 23 times a year, plus an annual Supplement on the IMF and an annual index. Opinions and materials in the IMF Survey do not necessarily reflect official views of the IMF. Any maps used are for the convenience of readers, based on National Geographic’s Atlas of the World, Sixth Edition; the denominations used and the boundaries shown do not imply any judgment by the IMF on the legal status of any territory or any endorsement or acceptance of such boundaries. Material from the IMF Survey may be reprinted, with due credit given. Address editorial correspondence to Current Publications Division, Room IS7-1100, IMF, Washington, DC 20431 U.S.A. Tel.: (202) 623-8585; or e-mail any comments to