The Rise And Fall of American Growth by Robert Gordon • Princeton University Press • 2016 • 784 pages • $39.95
If there’s one thing we all agree on in this cantankerous country, it’s that we live in an age of dazzling technological change, change so rapid we can barely keep up. Remember the iPod? Remember how amazing it seemed at first? And then remember how, in just a few years, it became completely obsolete? That’s just one of the many examples that have us all convinced that, with respect to technological advancement, surely we live in the most astonishing time ever.
Erik Brynjolfsson and Andrew McAfee in The Second Machine Age project that the pace of technological change will continue at a high rate. In particular, they and other technological optimists believe that advances in artificial intelligence will open up new possibilities across all industries, leading to better health and smarter, interactive products. The only problem will be that automation and the use of robots in homes, offices, factories, and roads will become so effective that more than 30 percent of the labor force will be unable to find jobs.
But Robert Gordon thinks otherwise. In his magisterial new work, The Rise and Fall of American Growth, the influential economist “distinguishes between the pace of innovation and the impact of innovation.” (emphasis in original.) He argues that the main benefits of the digital age of information and communication technologies (the “third industrial revolution”) had already occurred by 2004. Further, he says, the new gadgets improve our recreation and leisure but impact only in small ways our consumption of food, housing, health, clothing, and transportation. Finally, Gordon doesn’t see major advances in the future of health care, or gains from the use of small robots, 3-D printing, big data, or driverless vehicles. All this adds up to a bleak forecast on growth: Gordon says growth in the United States will average about 1.3 percent per year for the next 25 years, which is considerably less than the 2 percent growth that existed over the twentieth century.
Gordon asserts that our most muscular period of economic growth ran from 1870 to 1970. In painstaking detail, he shows how people’s lives were changed by the confluence of many advances that improved life expectancy and living conditions dramatically. These innovations, he writes, could only happen once, and they brought improvements to people’s lives that were even better than reflected in the official growth rates of real GDP per person. His backstories of many of the innovations make the book a fascinating read for anyone who wants to understand what the economy really does.
For example, after 1970, Gordon sees economic change as being more “evolutionary” than “revolutionary.” By the time the third industrial revolution started in 1980, most of the advances in living standards were already in place. And Gordon doesn’t foresee the possibility of any major advances in living conditions beyond how upper middle-class Americans already live today. Since he doesn’t see this upper middle class expanding to absorb a greater proportion of the population, he is left to make the negative prediction that real incomes in the middle and bottom of the income distribution will rise at the anemic rate of 0.3 percent per year for the next 25 years.
While he never discusses or envisions the political consequences of such a future, it is not hard to imagine that the populist anger that fueled the election of Donald Trump will only grow in intensity if the gaps between the rich and the middle grow larger, while the incomes of the bottom 70 percent grow ever so slowly. If Gordon is right, the strains from the 2016 presidential campaign—in which both candidates were disliked and in which one candidate raised the possibility of diminishing our political liberties—will intensify and upset the political stability that has characterized America for so long.
Robert Gordon has been a leading macroeconomist for the last 40 years and in 1995 served on the Boskin Commission, the seminal economic panel that found that inflation was being overstated by at least 1 percent a year. He begins his book with the 1869 hammering of the Golden Spike, which symbolically united Americans from coast to coast. At this time, life expectancy was short (45 years). And as to living itself, well, here’s Gordon:
The typical 1870 household faced not just stark limitations on the quantities of consumer goods that it could afford, but also the difficulty of the work effort by both males and females. . . . The greatest hardship was the unrelenting toil of the rural and urban housewife, who not only had to carry out the daily physical task of moving clean and dirty water, coal, wood, and ashes, but also had to expend further labor making food and clothing. . . . Processed food was largely unavailable, and fresh meat was unsafe, so the diet was a monotonous succession of salted pork and starchy foods. . . . Clothing was crude and for most women home-made [without sewing machines]. . . . Dwelling units in 1870 universally lacked indoor plumbing, running water, waste disposal, electricity, telephone, and central heating. . . . Every family . . . was directly or indirectly dependent on the horse . . . [with] manure and urine distributed on the streets. . . . For the 75 percent of the American population classified as “rural,” the overwhelming impression from today’s perspective is isolation.
In addition, few went to school after the age of 12, and child labor was common. In terms of health care, the practice of medicine was primitive and hospitals were breeding grounds for disease. In 1870, the incidence of infant mortality was 170 per 1,000 live births. Water wasn’t chlorinated or filtered, leading to 37 percent of deaths in 1900 being attributable to infectious diseases. And industrial accidents led to the maiming and deaths of many thousands of people.
What followed was the “special century,” when a series of “once only” advances occurred. Again, here is Gordon:
By 1929, urban America was electrified and almost every urban dwelling was networked, connected to the outside world with electricity, natural gas, telephone, clean running water, and sewers. By 1929, the horse had almost vanished from urban streets, and the ratio of motor vehicles to the number of households reached 90 percent. By 1929, the household could enjoy entertainment options that were beyond the 1870 imagination, including phonograph music, radio, and motion pictures. . . . By 1929 infant mortality had almost been conquered. . . . By 1929, work days were shorter . . . and electrical appliances had begun to lighten the everyday chores of home production.
Gordon is excellent at dealing with the nitty-gritty. There are many detailed discussions of changing technologies and how they translated into an improved standard of living. He gives the backstories of many of the most important innovations and covers topics like changes in the distribution system, the rise of banking and financial services, and the importance of government investments and regulations that facilitated economic growth. For each innovation, he shows the rate of adoption growing quickly. As people demanded all of the five connections named in the quote above, old housing was demolished and new structures appeared. Another set of changes that improved living conditions include processed food and the growing use of refrigeration, which provided for much more variety in people’s food consumption—and made for a healthier diet, as fresh fruits and vegetables became more consistently available. This was aided by an organizational shift from local general stores with few options to national chains of supermarkets and mail-order catalogs.
These changes brought about enormous benefits. One indicator that Gordon tracks closely is life expectancy. He cites an alternative calculation of economic growth produced by William Nordhaus that adds to GDP growth the economic value of each additional year of life. Instead of a real growth from 1900 to 1950 of 2.1 percent per year, using the conventional GDP per capita accounting, Nordhaus’s “expanded” measure calculated a GDP growth of 4.2 percent per year over these same years. In other words, the growth in the American standard of living over these 50 years was underestimated by half.
The title of his short introduction to economic developments after World War II is “the shift from revolution to evolution.” His perspective shifts; he no longer emphasizes nonmarket gains that positively affect people’s standards of living because he thinks that almost all the gains in living standards are now incorporated well in productivity gains as measured in our GDP accounting.
In the last quarter of his golden century (1945-1970), there was the consolidation and spin-offs of previous innovations, especially in air conditioning, air travel, television, and the construction of the interstate highway system. Given the distortions of ten years of the Great Depression and five years of world war, there was plenty of pent-up demand for new housing, more children, new suburbs, and more roads. Consequently, the growth rate remained high.
It should be noted that the ability to measure growth began in the 1930s, and growth became widely measured in various countries only after World War II. All historic measurements are based on applying modern methods to fragments of historical data that are available. The key elements of growth are based on the market transactions of selling to consumers and business, hiring labor, and deploying machinery. Although we think that “productivity” is based on the application of new organization and scientific methods, there is no direct way to measure this. Thus, “total factor productivity” is the residual from the increase in final output minus the increase of the amount and quality of physical capital and labor.
From 1970 to 2010, the growth of total factor productivity declined by one-third. According to Gordon, this was no accident. For him, the advances made after 1970 did not spur multiple spin-offs or gains in the quality of life as big as the ones that occurred during the golden century. Advances in jet engines and safer and more fuel-efficient automobiles are portrayed as just small steps forward. Similarly, in health care, the discoveries of penicillin, other antibiotics, vaccines, and greater diagnostic detail (X-rays, CT scans, and MRIs) drove life expectancy up by eight years between 1950 and 2008—which compares unfavorably with the 35-year gain between 1870 and 1950.
His treatment of the effect of the third industrial revolution based on improvements in microprocessing emphasizes that their impacts were mainly limited to communication and recreation, but not much more. He dutifully shows that the adoption of computers and smartphones increased at an astonishingly high rate. This was caused by the massive advances in the number of processes that could be completed by ever more powerful microchips. In essence, we were getting a better product with more capabilities year after year at the same or lower prices. Yet Gordon doesn’t see these gains as being translated into broader improvements in living conditions. He sees them as limited to improving communication and expanding our entertainment options.
To back this claim up, Gordon reminds the reader of the slow reported productivity growth of these years and minor improvements in living standards of people in the middle and lower classes. His pessimistic comments on the state of the middle class are numerous:
- “the bottom half of the income distribution has begun to drop out of the middle class”
- “40 percent of college graduates have been unable to find jobs that require a college education”
- “49 percent of middle-class workers will be poor or near poor during retirement”
- “The average real income in the bottom 90 percent was actually lower in 2013 than it was in 1972″
All of these statements are based on other people’s research and have been widely cited. Yet Gordon also presents results from the Congressional Budget Office (CBO) study on household incomes that shows yearly middle-class growth of after-tax income to be 1.6 percent a year between 1979 and 2011—which translates into a real gain of more than 50 percent. These data are included to show another approach to measuring changes in standards of living and there is no discussion of why these numbers are so different from the negative findings cited above.
While I cite the CBO data in my book Rebound to refute the dire findings cited above, Gordon presents both negative and positive assessments of middle-class income growth even though he praises the CBO approach as being the “most comprehensive” and “superior to the other data sources.” It should be noted that the 2013 difference between income data sources is huge: The median income of Piketty and Saez’s tax filers was $29,000, versus a median household income of $54,000 in the Current Population Survey, and median income of more than $70,000 in the CBO.
In his final chapter, Gordon presents his most newsworthy and controversial finding: that middle-class income growth will be miniscule over the next 25 years. Projecting the future growth of middle-class income is a dicey proposition, and Gordon’s approach should definitely be thought of as just one possible “guesstimate.” But his numbers should also be considered the low end of what might happen, because he consistently uses the most pessimistic assumptions to arrive at his projection that middle-class growth is going to be only 0.3 per year from 2015 to 2040. Here’s how he gets there, with my comments of what he misses:
- Instead of using the 1.6 increase per year in labor productivity from 1970 to 2014, he excludes the strong growth years of 1994-2004 because he views them as anomalous and projects future growth at 1.3 percent per year. This is based on the unrealistic assumption that all of the digital advances in the future will have very little effect on improving people’s living standards. For example, major technology firms are investing billions of dollars in artificial intelligence and robotics. While it is reasonable to say that this won’t lead to massive technological unemployment as some of the “techno-optimists” believe, it is hard to see that there won’t be some positive effects from all of these efforts.
- He thinks that the “quality” of labor will no longer increase and subtracts an extra 0.3 percent per year because he believes that there will be little increase in educational attainment in the future labor force (which contrasts with the large increase in the share of those with post-secondary education from 1950 to 2000). This underestimates the recent gains in college-going and rules out any potential success of the stated goal of increasing the share of youth and young adults with post-secondary training. In addition, interactive computer learning may become more prevalent in terms of learning specific skills that are closely related to workplace needs.
- In making the transition from output per worker to output per person, Gordon subtracts another 0.2 percentage points to get to 0.8 percent growth per year because of the decline in hours worked per person, owing to the retirement of the baby boomers. This violates Gordon’s premise that we are interested in people’s standard of living and not just their market incomes. As discussed below, people being freed from having to work should be considered a positive in terms of living standards.
- He reduces this figure further to 0.4 percent due to the effect of the wealthiest hogging income growth leaving less for those in the middle of the income distribution. This assumes that income inequality will continue to increase from its current high levels to levels never reached before in American history. Since Gordon believes that the relevant income measure is after-tax income, I argue below that the political pressures of ever-increasing inequality may lead to tax and income policies that will reduce after-tax inequality.
- He subtracts a final 0.1 percent per year due to a technical factor relating to the difference between GDP growth and personal income growth to get to his 0.3 percent yearly growth in middle-class standard of living. (This factor is too small and too technical to be worthy of comment.)
So the person who, at first, went to great lengths to show that increases in the standards of living were greater than the official increases in GDP for the years 1870 to 1970 is also the one who makes the most (plausible) negative assumptions to project slow middle-class growth in the future. An easier approach might have been to start with the CBO yearly real growth of about 1.4 percent during the years of slow productivity growth. Even if one decided to adjust this down a few tenths of a percent due to factors cited by Gordon, the projected growth rate would be at least 1 percent per year.
There are three other factors that might affect growth in the middle positively. First, a group of eight researchers from the Bureau of Labor Statistics and the Bureau of Economic Analysis released a working document in 2011 that argued that official statistics had vastly overestimated the rate of medical cost inflation and that, in reality, there was much more real growth in output than officially reported. Nordhaus, to return to his work, estimates that real GDP adjusted for health gains from 1950 to 2000 was 3.5 percent per year versus the conventional GDP growth rate of 2.1 percent. It is very likely that gains in health care will be one of the brighter spots in our future.
Second, there is the issue of working less. Both Keynes and Galbraith saw that the compounding of economic growth would lead to a situation in which we run out of new things to consume. Their answer was that part of growth in the future would come from a reduction in the number of hours worked. In fact, we have done this: Part of living longer and part of going to school longer means that people work during a smaller fraction of their lives. In my book Rebound: Why America Will Emerge Stronger from the Financial Crisis, I showed that the share of one’s life spent working for pay declined from 70 percent in 1950 to 52 percent in 2010.
These added years of not having to work full-time yet being able to live at least modestly well for most college students and retirees represent a significant increase in the time available to pursue recreation and more social interactions. Although there is inequality among the elderly, it is one of our great public policy triumphs that seniors in America today live as well as they do while having access to very subsidized medical care. It is not a perfect system, but it is much better than the alternatives that they had in the past. For example, the elderly poverty rate in 1959 was 35 percent (more than 10 percentage points higher than the non-elderly); in 2015, the elderly poverty rate was 8.8 percent, which was 5 percentage points lower than the non-elderly rate.
Third, Gordon does not account for any political response to the continued rise of income inequality. The recent election showed us that people are upset, to put it mildly, with our political leadership. Many were excited by Bernie Sanders’s message of taxing the rich, and others were attracted to Trump’s vows to revive manufacturing, increase coal production, and so on. If the future leads to greater divides between the top 30 percent and the bottom 70 percent, the attraction of left and right answers will grow—and result in very different potential trajectories, depending on which side wins.
This hefty 750-page book covers the ups and downs of economic growth in the United States over the last 145 years as well as the great inventions that radically improved people’s quality of life during that time period. For the last 30 years, Gordon has also championed the idea that our price adjustments overstated inflation and understated the real gains that positively affected middle-class Americans. But when it comes to projecting the future, Gordon seems committed to making a negative, pessimistic statement.
While the book has gotten attention because of its bold projection of slow growth in the future, this is actually just one small element of a magnificent and detailed presentation of how our economy has changed since 1870. Most people don’t fully appreciate what life was like in the past and Gordon gives a blow-by-blow description of how people lived in America from 1870 on. In addition, he carefully explains how each new innovation was created and how its adoption changed people’s lives.