F rom the early-nineteenth-century introduction of steam power through the dawning of the age of the microchip in the post-World War II era, real economic growth in America averaged 3.8 percent per year. That meant economic output doubled roughly every 19 years. Then after the 1970s, growth collapsed. During the 1980s, growth averaged just 2.7 percent per year and since 1989 only about 2 percent. At that rate, it will take nearly 36 years for gross domestic product (GDP) to double again.
Despite this performance at well below historical trend rates, many mainstream economists and the journalists who follow them have concluded that the new lower growth rate is inevitable and more or less permanent. Economist Paul Krugman suggests that we now live in an "age of diminished expectations" and we had better get used to it. The Council of Economic Advisers (CEA) forecasts 2.3 percent growth through 2002. Generally, both the Federal Reserve Board and the Congressional Budget Office agree with this assessment. Popularizing this theme, Jeffrey Madrick's 1995 book concluded that we had reached The End of Affluence.
This pessimism about future growth rates is predicated on unassailable arithmetic but, as we shall see, questionable assumptions. Mathematically, the rate of economic growth cannot exceed the rate of labor force growth plus the rate of growth in labor productivity-that is, total hours worked times output per hour worked. This tautology sets the speed limit on how fast an economy can grow; if we exceed the economy's growth capacity, we will reap only inflation.
Typical forecasts of economic growth-the CEA's, for example-predict that the labor force will grow at about 1.2 percent per year at least through 2002 while productivity growth will creep along at about 1.2 percent a year as well. Adding these together (and subtracting a tenth of a point for the expected slowdown in growth due to smaller government outlays and a smaller farm sector) yields the "official" 2.3 percent growth rate forecast that pretty much everyone has come to accept as the most likely future scenario. Optimists predict that with a little luck we could push it to 2.5 percent. The bridge to the twenty-first century apparently is being built without a high-speed lane.
Small differences in growth rates over a sustained period yield huge differences in our standard of living. Whether the economy can grow at 2.3 percent or, say, 3 percent a year may seem a quibble, but that annual seven-tenths of a percent is hardly trivial. Between now and 2007, the total difference between these two rates is $3.1 trillion worth of GDP-an average of more than $300 billion a year. That $3.1 trillion could solve the Social Security "crisis," deal with the federal deficit, and represent a large down payment on rebuilding the cities and cleaning up the environment-not to mention what it could mean for employment, wages, and family incomes. Misreading the economy's speed limit has its costs. This is particularly true if investors temper their own expectations of growth based on "official" forecasts and limit their investments in new technology and in new plants and equipment believing that any larger investment will simply leave their capital stock idle in the face of slow-growing demand.
It may seem prudent to "diminish our expectations," but there is the distinct danger that limiting our expectations may become a self-fulfilling prophecy. In this article, we suggest why. Instead of forecasting growth rates, the Federal Reserve Board, the White House, the Congressional Budget Office, and the economists on whom these bodies rely for advice may be inadvertently setting growth rates. It would surely add credibility to the economics profession if its slow-growth forecast turned out to be accurate-but it would not be very good for everyone else!
WHO WORKS, AND WHY
We think the economy can grow faster-provided the growth prospects are not sabotaged. There is considerable evidence that the mainstream forecasts of both labor supply and output per worker are too pessimistic, given emerging underlying forces in the economy. The supply of labor and the potential growth of productivity are both more elastic than the standard view admits. Moreover, the conventional wisdom relies too heavily on an outmoded understanding of how technology and growth are connected.
Consider labor supply. We may not like all of the social consequences of an increasingly "overworked" America, but more Americans are working, and they are working longer hours. We are finding that the labor force is not tapped out, but is ready to work more when jobs are there. The economist's old homily, Say's Law, which says that "supply creates its own demand," has been stood on its head. Now, increased demand for labor coaxes out a new source of labor supply that can contribute to faster growth. This ups the economy's speed limit a notch.
The increased labor supply shows up in two forms. One is an increase in the labor force participation rate-the fraction of the population that works or seeks work. [See "The End of Unemployment?".] After growing for decades, the labor force participation rate flattened out in the late 1980s and early 1990s. Economists and demographers took this to mean that with women's participation reaching a peak and older workers retiring earlier, the overall labor force participation rate had reached a plateau. Between 1989 and 1994, the rate was essentially flat at 66.6 percent. But lo and behold, since 1994 the rate has been on the rise again. By March of this year, the rate was up to 67.3 percent-adding 1.4 million additional workers to the labor pool.
This resurgence in part reflects an apparent end to four decades of decline in labor force participation rates among workers over age 55. With improvements in health and changes in law that permit older workers to keep more of their Social Security income while continuing to work, more older Americans are choosing to remain employed in either full-time or part-time jobs. Moreover, it appears that younger men who had "disappeared" from the official labor force-as many as six million according to Lester Thurow ["The Crusade That's Killing Prosperity," TAP, March-April 1996]-are beginning to reenter in response to better job opportunities. In other words, faster growth and more job opportunities coax out more workers; labor supply is not fixed by demographics. Extra labor supply from these sources will offset the projected decline in 25- to 34-year-olds over the next decade, the legacy of the baby bust generation. Thus, assuming adequate labor demand-something that macroeconomic policies at home and coordinated negotiations among countries can promote-the "graying of America" need not mean a decline in the growth rate in labor force participation.
Even more important is the growth in hours worked per worker-a factor that goes undetected in the official labor force participation and employment statistics, since all workers who work at least one hour a week are simply counted as employed. From 1967 through 1982, the average annual hours worked by prime-age workers (age 25-54) declined from about 1,975 to about 1,840. This was primarily the result of women entering the labor force in part-time positions. Since 1982, however, the trend has reversed, so that by 1995 the average work year was back above the 1967 level. Just since 1991, with the expansion of the economy, average hours worked per employee have increased by nearly 3 percent. That is the equivalent of adding 3 percent more workers to the American labor force if the average work effort had remained unchanged at its 1991 level.
That we have been able to drive the unemployment rate down below 5 percent without igniting wage-led inflation is testimony to the fact that there is a good deal of labor supply in the pipeline when labor demand exists to employ it. Because of its construction, the official unemployment rate simply fails to capture it. Of course, at some point the typical workweek and the fraction of the population that works reaches a natural limit. But given healthier senior citizens, women's equal participation in the labor force, welfare-to-work efforts, and training programs for the conventionally unemployable, labor supply should grow faster than population for some time.
A PRODUCTIVITY RENAISSANCE
The other component of growth-labor productivity-has been the chief culprit in the growth slowdown since the early 1970s. From 1870 through 1973, productivity increased by an average rate of 2.4 percent per year. In the immediate post-World War II era, productivity was absolutely booming-growing more than 3 percent a year. After 1973, productivity growth totally collapsed, for reasons that most economists consider a mystery. (The OPEC oil shock, the demise of the Bretton Woods system, and the high-interest-rate austerity that followed are considered possible suspects.) For a quarter century, productivity has been growing at barely 1 percent a year-a pace even worse than that of the Great Depression. The official projections for economic growth are based on a continuation of this dismal record.
Yet there is good reason to believe that we are on the verge of a productivity renaissance. This is already evident in the manufacturing sector, where productivity growth rates are back up to the levels enjoyed during the postwar glory days. The recent drag on productivity has all been in the service sector of the economy, where most of the growth in the economy is now centered. But even here there is now evidence of a turnaround. From 1989 through 1995, productivity growth for the entire nonfarm business sector of the economy averaged only 0.9 percent per year. Since then, the rate has averaged 1.3 percent-a sizeable improvement. Longer-term historical data suggest that the productivity slump bottomed out in the late 1980s.
M ore important perhaps than the actual recent numbers is the growing awareness that overall productivity growth has a history of long cycles based on the introduction of new technologies. This helps to explain the productivity paradox of the information revolution. The work of such economic historians as Paul David, Nathan Rosenberg, Luc Soete, and Chris Freeman, shows that in every new technological era-be it led by the introduction of water power, steam power, or electricity-there has been a lag before productivity surged. The decline in productivity growth rates often lasts two or three decades, the time it has taken for each new technology to be "debugged" and diffused.
The computer revolution very likely follows the pattern. If so, we likely have lived through the downside of a technological transformation and we are just about to receive its benefits. For nearly two decades now, we have seen the rapid introduction of new hardware and software as the technology revolution moves toward greater maturity. Every time we have begun to become proficient with new technology, newer hardware and software promises even greater productivity. But it takes time to learn, and during this time productivity growth is flat or even declines. For those who have kept up with the innovations in computer operating systems, think of the time and effort "wasted" in moving from DOS to Windows 3.1 to Windows 95. Just as you were moving up the "learning curve" with one operating system, another one came along and you had to pause while you moved onto the next learning curve. Over the long run, your productivity improved-or will improve-but the process is filled with fits and starts.
Moreover-and here is where mainstream theory really misses the boat-Paul Romer, Richard Nelson, and others have shown persuasively that the productivity dividend from the introduction of any new technological paradigm is fully realized only when all of its complements are in place. The hardware has to work with the software. The skills of the workforce need to be upgraded to utilize it. Old managerial routines that stand in the way have to be replaced. This all takes time-and now appears to be well along. A growing proportion of the workforce is now computer literate. Investments in training are beginning to pay off as more and more workers report that they are using computers and related equipment on the job. The human process of learning by doing is now increasingly routinized.
We have already seen this process flourish in the manufacturing sector. With greater emphasis on user-friendliness, and with the accelerated diffusion of the new easier-to-use technologies in the service sector as well, the productivity promise of the computer and related information-processing equipment is now on the close horizon. This, we believe, helps to explain the bottoming out of the productivity slump and heralds strong positive growth in the near future.
FEEDBACK LOOPS AND SELF-FULFILLING PROPHESIES
Our estimates based on recent trends suggest that we can claim another 0.3 to 0.4 percent growth annually in the labor force over the next decade and another 0.3 to 0.4 percent growth per year in productivity. This means that there is clearly room for the economy to meet a 3 percent annual growth target-the amount necessary to garner that $3.1-trillion GDP bonus.
But we surely will fall short of that goal if public and private policies sabotage growth. Here there is much to unlearn from history. Though the Federal Reserve Board has lately allowed unemployment rates to slip below what it once considered a natural floor, there remains a powerful monetary policy bias against faster growth [see "
Test the Limit
," James K. Galbraith].
I n the stagflation of the 1970s, many economists became convinced that the natural rate of unemployment was an immutable law. But in the 1980s and 1990s, the real trade-off between inflation and unemployment has become more benign due to increased actual and potential competition as a result of increased world trade, the proliferation of offshore production sites (and producers), new technology, changes in the nature of labor markets, and industry deregulation. In fact, the recovery since 1992 has seen the unemployment rate fall from 7.5 percent to 4.9 percent without any increase in inflation whatsoever-a clear case where the facts inconveniently trump received theory. The Fed is nonetheless poised to raise interest rates in the near future if unemployment continues to fall or simply remains below 5 percent.
Such action would not only sabotage growth in the short run; it compromises the potential for long-term growth because of an important feedback loop between monetary policy, the investment climate, and future growth. For private business to continually invest in new capital and new technologies, firms must believe that the output created by the new equipment will actually be sold. Here is another case where Say's Law operates in reverse. If private business believes demand growth will be slowed in the future, it will reduce its current investments accordingly-producing the self-fulfilling prophecy of slow growth. If it expects, on the other hand, that growth will be encouraged by policymakers, then business will more likely make the investments that will feed back in the form of higher productivity and faster growth. The prospect of more demand for products produces the supply of investment and innovation needed to make that prospect come true. If the monetary authorities signal that 2.3 or 2.5 percent growth is the most we can achieve, then that is what we are going to get.
The demise of fiscal policy in the service of balancing the budget at any cost also sabotages potential growth. When deficits continue to rise without limit as a percentage of GDP, there is no doubt a drag on economic growth. But targeting a zero deficit is not necessarily good for growth either. Economic expansion in a technological age requires continuous investment in public infrastructure, in generic R&D, and in training and education. If we continue to sacrifice these on the altar of budget balance, we could undermine the very complementarities that growth requires. A strong case can be made that we should now be spending more on public investment, not less, if we do not want to undermine the prospects for growth.
Industry too must consider all of the complements to growth that need to be in place to assure a new era of economic expansion. Continuing to build closer working relationships with labor where both workers and managers are committed to productivity, quality, and innovation is one area where much is still to be accomplished. We must leave behind an economy where workers are treated as expendable costs rather than the crucial assets they are or could be. Expanding the level of employee training throughout the enterprise, rather than (as at present) mainly at the top, is another area where good business practice could enhance national economic growth.
THE CASE FOR OPTIMISM
All of this suggests that faster growth is possible-if we don't sabotage it. Changes in the labor supply are providing one leg. The maturing of the information age is providing the other. If we can make sure that fiscal and monetary policy do not sabotage growth and if we can encourage businesses to expand their investments in human capital to meet the investments they have already made in their physical plant and equipment, we will be on the road to faster growth. If we use that growth dividend wisely, we can raise living standards, reduce the gap between the rich and the poor, and help solve many of the pressing social problems we face. We can repair many of the gaps in the social safety net for both those who can work and those who cannot.
We may not be able to quite reach the pinnacle of economic growth we enjoyed during the post-war glory days, but we surely can do much better than the growth depression we have endured for the past quarter century. Indeed, we might have begun down this road a decade ago and avoided a great deal of economic and social pain, if we had shed old resistances earlier and adopted the appropriate pro-growth policies already.
