In 1990 Stanford economist Paul Krugman wrote a superb primer on U.S. economic policymaking in the 1990s that he called The Age of Diminished Expectations. In it he argued that the most important problem facing the United States was a slowdown in long-term economic growth. As usual Krugman got his economics right. But his description of the mood of America in the 1990s was wrong. Instead of bringing their expectations into line with the economy's lower productive potential, Americans have been in continual denial, lashing out at foreign competition, immigrants, the poor, minorities, declining values, and above all, big government as the source of their problems.
In The End of Affluence, Jeffrey Madrick follows in Krugman's footsteps with a lucid, well-documented portrayal of America's economic plight. Madrick's central thesis is that America has entered a new era of slow growth that requires a radical change in mindset. While his message is not original, it certainly warrants repetition, since so few Americans have taken the implications of slow growth to heart, and no one has yet come up with a complete explanation for why it has happened and a full prescription for what should be done about it. Madrick's book is interesting chiefly for its description of the disease. But his analysis of the disease's origins is faulty, and his prescriptions are weak.
THE TWO-PERCENT DISSOLUTION
For over a century after the Civil War, the U.S. economy grew at a yearly rate of 3.4 percent. For the past two decades, by contrast, the pace has been almost a full percentage point slower. For over a century, growth was driven by output per worker that was rising two percent annually, enabling real wages to grow at a similar rate. This "two-percent solution" was the key to the American Dream. The miracle of compound interest resulted in wages that doubled every 35 years, and so, for over a century, each generation lived twice as well as its predecessor.
As Madrick notes, the achievement of the dream not only brought prosperity, it also shaped attitudes. It both reflected and reinforced Americans' optimistic, individualistic, self-reliant ideology. The achievement also shaped Americans' attitudes toward the poor. In an economy of opportunity, which rewards diligence and hard work, the belief went, those who fail have only themselves to blame. By contrast, in European economies with longer traditions of shared misery, solidarity with the less fortunate became more common.
But since the early 1970s, America has been rudely awakened from its dream. The pace of labor productivity growth has been cut by more than half. Between 1973 and 1994 the average growth in output per worker in the business sector was less than one percent, and reflecting this slow productivity growth, wage growth has slowed to a crawl. Had the long-run growth rate in real compensation of two percent per year been maintained since 1973, average compensation today would be higher by more than a third.
Moreover, the aggregate data disguised a second noteworthy development: growing inequality. Only families in the top fifth have experienced any income growth at all. Similarly, college-educated workers have been the beneficiaries of all the wage growth that has occurred; wages of workers with a high school education or less have declined outright.
Madrick does a convincing job of documenting these changes and drawing out some of their effects on policy. Had the economy grown just one percent per year faster over the past two decades, the recent feuds over health care and the budget would not have taken place. All Americans could easily have been provided health care. Given actual spending outlays, federal tax revenues would have been sufficient to cut the national debt in half. Or, had Americans chosen to, they could have had both more spending and tax cuts. Madrick also points out that despite these changes, the attitudes of earlier times prevail. Americans retain their optimism and, mercilessly rejecting politicians speaking of malaise, reward those proclaiming "morning in America." They see no reason to provide more help to the poor even in an environment with diminished opportunities. Indeed, they are taking steps to end welfare as we know it in an effort to foster attitudes of self-reliance.
What explains the end of affluence? According to Madrick, the causes are technological change and international competition. The golden era of American growth was driven by mass production in a self-sufficient economy with large internal markets. The new era is based on flexible production methods in an economy with fragmented markets. Under the old system, sustained and predictable gains in productivity growth came from applying routinized mass production methods on an ever-increasing scale. This evolution provided a secure environment that supported large firms prepared to invest in capacity and hire unskilled workers for jobs that were stable and relatively well paid. In the new era, however, economies of scale are less important. The capacity to produce flexibly allows for increasing product and service differentiation. It provides opportunities for smaller firms to fill niche markets and reduces the competitive advantages of large firms.
In Madrick's view, these changes have been damaging. Ironically, the proliferation in variety raises costs in the long run. The lack of standardization leads to greater volatility and uncertainty, resulting in lower levels of investment, more unused capacity, smaller gains in productivity, and less-secure jobs that require greater levels of skill. But Madrick never convincingly explains why America has been beguiled into introducing flexible production systems when they actually make it worse off. Why are consumers misled into paying more for exercising their taste for variety, when they should stick to cheaper, standardized products? Why, to use an example Madrick cites, do viewers misguidedly pay for cable tv, which provides hundreds of channels, instead of confining their attention to the big three networks whose programs are available at no charge? And why have producers responded to these demands when, according to Madrick, they could have achieved higher productivity growth and lower costs by remaining large and producing more standardized products?
Madrick's explanations require one to abandon the presumption that consumers are the best judges of what makes them better off when they demand more variety, and that producers are the best judges of how to earn profits when they take advantage of flexible production methods to meet these demands. It is far more plausible, however, that consumers and producers really do know what makes them better off--that the gains brought about by the new production modes have been much greater than could have been achieved with the old, and that the slowdown in the economy-wide growth rate has occurred despite, not because of, flexible production systems.
THE MANUFACTURING MODEL
Madrick argues that declining international competitiveness has compounded the change in technology. America's global lead rested on mass production made possible by its large, integrated internal market. But this advantage was diminished, first, when declining trade barriers and transportation costs allowed smaller countries to enjoy similar benefits from the world market, and second, when flexible production systems made scale less important. By the 1970s, therefore, America lost its leadership in mass production industries, and according to Madrick, "The resulting loss of markets meant fewer jobs and lower wages for American workers, the abandonment of some industries, idle capacity in many others and reduced capital investment overall."
It is true that manufacturing's share of U.S. payroll employment has declined rapidly--from 26.2 percent in 1973 to 16 percent in 1994--and that one reason was the emergence of a trade deficit in manufactured goods. But even without the trade deficit the qualitative picture would have been pretty much the same. According to estimates by Jeffrey Sachs and Howard Shatz, for example, employment in manufacturing in 1990 would have been about 6 percent greater, or just over 18 percent of total employment.-1 Although Madrick acknowledges that trade remains a small part of the economy, he argues that internationalization created a number of powerful indirect effects, such as "rising levels of uncertainty, greater instability and business strategies to defend against foreign encroachment, which in turn reduced capital investment, reinforcing the declining trend in productivity and wage growth."
If Madrick's explanation for America's slow growth were valid, we would expect to see the largest slump in productivity growth in the manufacturing sector, where trade is very important, and the introduction of flexible production systems has proceeded most rapidly. Instead, according to the Bureau of Labor Statistics, the real laggards in productivity growth have been mining and construction (scarcely exemplars of flexible production); electricity, gas, and sanitary services; several sectors in finance; and most important, services. Madrick's preoccupation with manufacturing and international competitiveness lead him to forget that America's slow productivity growth has been concentrated in sectors that are almost totally sheltered from international competition.
It is surprising that despite his voracious consumption of the economics data and literature, Madrick fails to grasp the implications of this critically important fact: productivity growth in American manufacturing has been as rapid since 1980 as it was before 1973. Between 1979 and 1989, for example, according to the Bureau of Labor Statistics, the 2.7 percent annual rise in output per work-hour in U.S. manufacturing was exactly the same as the rate between 1955 and 1973.-3 After a period in the second half of the 1970s and the early 1980s in which it slumped, American manufacturing productivity advanced just as rapidly during this era of flexible manufacturing and increased international competition as it did during the golden era of mass production.
In fact, America's manufacturing performance did not contribute to the slow rise in American productivity since 1980. On the contrary, the performance in manufacturing was the major reason economy-wide productivity grew at all. Outside manufacturing, the average annual increase in output per worker was barely half a percent!< The primary reason for the declining share in manufacturing employment in the United States is relatively rapid productivity growth, not foreign trade. Just as the shrinking share of employment on farms resulted from improvements in farm productivity in earlier times, so manufacturing's declining share of employment reflects its relatively more rapid productivity improvements. In particular, over the past 15 years, the difference between productivity growth in manufacturing and productivity growth in the rest of the economy has increased markedly.
Madrick ascribes recent U.S. trade performance to the weak dollar. He overlooks that the dollar is actually no weaker today than it was in the late 1970s.> By contrast with the 1970s, over the past 15 years America's international buying power has not been depressed by a falling dollar and import prices have increased no faster than American prices. Madrick neglects recent evidence that American manufacturers have regained their leads in high-tech sectors such as microprocessors and dramatically improved their performance in basic industries such as steel and automobiles precisely because they have restructured, innovated, and adopted more flexible production methods. Indeed, America's rapid transition to the industries of the future in software, telecommunications, and finance is the envy of the world.
Moreover, the problem of slower growth is not unique to America. Compared to other major industrial economies, when it comes to prospects for growth, America is actually doing about as well (or as poorly) as Europe and Japan. While economy-wide productivity and real wage growth in Europe has been faster than in the United States, European employment growth has been much weaker. In the 15 years before 1990, Japan's growth performance was a stellar 4 percent per year. But in the face of the severe structural problems that emerged in the 1990s, estimates for Japan's long-term annual potential growth rate have been revised downward to between 2.5 and 3 percent. Thus, throughout the developed world, forecasters expect medium-term growth of between 2 and 3 percent because of slower labor force growth and productivity gains.
America has not competed poorly in recent years, but its growth problem remains. The U.S. manufacturing sector has improved its performance relative to other countries and is doing no worse in raising productivity than it did historically. Yet productivity performance in the economy as a whole is slow because manufacturing counts for less than 20 percent of total output. Slow growth had little to do with "competitiveness" in the first place. Productivity growth throughout the economy is what really counts.
Flexible production systems may well have contributed to America's growing wage inequality because they do appear to increase the demand for skilled and better-educated manufacturing workers. But they do not deserve the blame for slower productivity growth or recent trade performance. The real problem has been the failure of producers in services to apply these technologies in ways that raise productivity.
There are some who reject the claim that the productivity problem lies in the service sector on the grounds that productivity growth in services is hard to measure. But that has always been true. Despite the measurement problems, before 1973 measured productivity growth in the rest of the business sector was quite similar to productivity growth in manufacturing. To make the claim that slow growth is simply a matter of mismeasurement, it is not sufficient to argue that measurement is difficult; it would have had to become increasingly difficult since 1973. While this may be true in some services, it is unlikely that mismeasurement problems can fully explain this pervasive development.fi
There are others who claim that services productivity growth has now recovered. Restructuring of services will do for productivity growth in the 1990s what restructuring of manufacturing did for productivity growth in the 1980s. Certainly layoffs of white-collar workers in the services sector provide some support for this claim. But the evidence remains anecdotal and difficult to appraise because it is hard to distinguish a long-run trend from the normal fluctuations in productivity growth associated with the business cycle. There are some hints that productivity has partially recovered outside the manufacturing sector, but thus far serious studies find that the evidence is not compelling.-2 Although it is possible that those proclaiming a new era (or a return to the old era) are correct, there is as yet no firm statistical backing for their optimism.
THE DISMAL PRESCRIPTION
What then should be done? Madrick's preoccupation with international competition comes to the fore when he discusses how we might reverse the changes in the economy. He dismisses several policy measures not because they might not improve American productivity and income growth but because they might not lead to relative gains over other nations. He rejects making improvements in technology, for example, on the grounds that what we do today other nations will quickly copy tomorrow. But this objection confuses relative and absolute gains. I will surely be more fit if I train for and run the Boston Marathon, although I will certainly not win it. Even if others are able to derive more gains from new technologies than we are, it does not mean we would do better not to introduce them. Similarly, Madrick dismisses the potential to raise national welfare through trade with emerging markets because the country may run deficits. Madrick forgets that both sides gain from international trade. Even if we run trade deficits, it does not mean that trade makes us worse off. Indeed, Adam Smith pointed out that by buying abroad more cheaply what it costs more to make at home, we can raise our living standards.
As I have said, Madrick is stronger in describing the problem than in prescribing solutions. Yet in criticizing Madrick humility is in order because no one has yet solved the productivity slowdown mystery. It is something like the economists' version of cancer. We know there is a disease with very complex causes. We know the things you can do to increase your chances of a cure, but we cannot be certain of all the causes or the solutions. Madrick is right to warn us about misplaced optimism and the snake oil of quick fixes. But he is wrong when he discounts the potential contribution of measures such as improved education and training, better technology, higher savings and investment, and trade with emerging economies.
We do not know how to cure cancer, but we do know many things that improve the odds of survival: do not smoke, avoid direct sunlight, breathe clean air, drink clean water. Similarly, we do not know the cure for the slow growth in living standards but can be pretty sure that improved education, better technology, more public and private investment, and increased trade are likely to help. Nor should fears that foreigners may emulate or even do better than Americans inhibit the United States from taking measures that are in its own interest. America's prosperity rests not on being number one but on doing the best it can.
3 Multifactor productivity growth is a more inclusive measure of productivity, which takes account of inputs besides labor. By this measure, manufacturing productivity increased 1.4 percent annually in the 1980s, close to the 1.7 percent average between 1955 and 1973.
< According to Bureau of Labor Statistics estimates, between 1982 and 1990 multifactor productivity growth in the private business sector grew at an annual rate of 0.9 percent. Of this total, manufacturing contributed 1.2 percentage points, with the rest of the economy on balance dragging productivity growth down by 0.3 percentage points. Multifactor Productivity in U.S. Private Business, 1947-1990, Washington: U.S. Department of Labor, 1992.
> According to the Federal Reserve Board, the multilateral real trade-weighted value of the dollar (1973=100) stood at 85.8 in the fourth quarter of 1994 compared with 83.2 in 1979 and 84.9 in 1980. Economic Report of the President, Washington: GPO, 1995, p. 402.