6 CHAPTER 1
The (Even More) Overworked American
JULIET SCHOR
I consider Juliet Schor to be one of Americaâs intellectual treasuresâa scholar whose profound gifts have been devoted to making ours a happier and more balanced society. I first met her in 1991. Then an economist at Harvard, she was just finishing her powerful book, The Overworked American, the first to document and challenge the steady rise in hours worked by Americans since the late 1960s. Her book impressively examined the high price Americans are paying for their new epidemic of overwork, and it suggested a strong connection between long working hours and consumerismâwhat Schor called âthe work-and-spend cycle.â Schorâs work has been a wake-up call for many Americans, including myself, but sadly, the problems she analyzed have only grown worse, and are in even greater need of attention today. âJdG
One of the most striking features of American society is how much we work. Now the worldâs standout workaholic nation, America leads other industrial countries in terms of the proportion of the population holding jobs, the number of days spent on those jobs per year, and the hours worked per day. Taken together, these three variables yield a strikingly high measure of work hours per person and per labor force participant.
7 In 1996, average U.S. work hours surpassed those in Japan. And they havenât stopped climbing. Through booms and busts, both work hours and employment have continued to rise for more than three decades.
The Rise of Annual Work Hours
Just over ten years ago, I published a book entitled The Overworked American, in which I argued that contrary to the conventional belief that leisure time was increasing, U.S. working hours had begun an upward climb following the 1960s. My estimates caused a firestorm of controversy, but subsequent years confirmed the trend I identified. Work hours are indeed rising, and significantly so. And the trend has continued. Americans are now working even more than they did when The Overworked American was published.
The data I relied on were from the Current Population Survey (CPS) of the United States, a household survey. The Economic Policy Institute in Washington, which originally published my estimates, has continued to update them (see Table 1 for their latest calculations). What the data show is that from 1973 to 2000, the average American worker added an additional 199 hours to his or her annual scheduleâor nearly five additional weeks of work per year (assuming a 40 hour workweek).
Since the 1980s, work hours have risen steadily by about half a percent per year, a reality attributed both to the fact that weekly hours have gone up (about a tenth8 of a percentage point a year), and that people are working more days and weeks each year.
TABLE 1 / Annual Hours in the United States, 1967â2000
Viewed from the perspective of the household, which incorporates the rise in the participation of mothers in the labor force, the added burden of work has been even greater. Among all married-couple households, with heads of households in the 25â54 age range, total annual hours of paid work by both husbands and wives rose by a whopping 388 between 1979 and 2000, a gain of nearly 12 percent.
The increase has been even larger for some subgroups. Among those in the mid-point of the income distribution (the famously âsqueezed middle classâ) the average increase in hours worked annually was 660 per year, a rise of just over 20 percent.
The Controversy about Time-use Trends
As mentioned earlier, some researchers challenged my findings, most notably Thomas Juster, Frank Stafford, Geoffrey Godbey, and John Robinson. They all believed that Americans were actually gaining leisure time at a rapid clip. They based their conclusions on a different type of dataâdaily time diaries in which survey participants recorded their activities in fifteen minute time blocks.
My source of data, the Current Population Survey (CPS), was a large, representative sample of households. Respondents gave retrospective estimates of how many hours they had worked in the previous week. The time-diary researchers believed that people were over-estimating their work time in the CPS data.1
Some of the claims of the time diary researchers were easy to refute. Juster and Stafford, for example, argued against my conclusions on the basis of data which were already a decade out of date (ending in 1981), and which missed the large work-time increases of the 1980s. Similarly, time diaries do not measure annual hours, but only weekly ones. Given that the larger part of the increase in9 annual hours occurred because people were working more days per year, their emphasis on weekly estimates was misleading.
Another limitation of the time diary research was that it has never taken into account the substantial influence on hours of work that comes from variations in the rate of unemployment or the stage of the business cycle. Time-diary researchers compared hours at the peak of business expansions (longer) with hours in the midst of recessions (shorter). My methods corrected for all these macroeconomic influences.
Finally, the time-diary samples have been much smaller and unrepresentative of the whole country in ways which bias the results. One important virtue of the CPS is that it is a very large, representative survey.
On the other hand, time-diary researchers did have an important point. Diary data is superior to recall data, and their claims that people overestimate their working hours may be true. However, the issue under debate was less the actual amount of work-time than trends. As long as the tendency to overestimate is stable, the upward trend of the CPS data is still a valid indicator.
Furthermore, some researchers have argued that the general claim of an âoverworked Americanâ obscures important differences in experience by education and income level. Yet my original research found that virtually all subgroups in the labor force experienced an increase in hours, with the exception of the partially-unemployed.
Eventually, the controversy died down. The ongoing estimates of the Economic Policy Institute, as well as estimates provided by other economists, supported the finding that work hours were increasing. And by the mid-1990s, the change was recognized even in the time diaries. In the second edition of their book, Time for Life, Robinson and Godbey reported that their additional data collection efforts during the â90s were yielding a new trend: the number of hours that women worked each week had begun to rise dramatically.
The Ironic Effects of Laborsaving Technologies
Of course, there is a certain irony in all the work that Americans are doing. The U.S. led the world in the technological revolution of the 1990s, as the Internet, computers, wireless, bio-informatics, and science were supposed to yield stupendous productivity gains that delivered us from excessive labor. This was both a promise and a prediction. Consider Jeremy Rifkinâs book, The End of Work, which predicted that widespread technological change would increasingly make human labor superfluous.
As it turns out, however, the labor requirements of technology have very little to do with how many jobs an economy generates or how long people work at those jobs. Indeed, the first Industrial Revolution of the nineteenth century and the10 Technological Revolution of the late twentieth century teach us an important lesson: the introduction of labor-saving technologies are frequently the impetus for massive increases in work.
What accounts for this paradox? On the one hand, the new technologies of the late twentieth and early twenty-first century provided new opportunities for making money. As firms seized those opportunities, they required long work hours from their employees, especially in the high-tech sectors, and in better paid manufacturing industries, such as auto or steel, where heavy overtime became a permanent feature of life.
Employers were able to elicit those extra hours because structural changes in the labor market made it hard for people to resist. There were far fewer unions, while part-time, contingent, and temporary work had become more prevalent. Even at the end of the 1990s boom, Americans felt more insecurity about their job status than in previous decades. Finding a full-time job with good security, benefits, and promotional possibilities had gotten harder and harder over time. Landing one of those plums meant that long hours came with it.
At the same time, the booming economy reinforced a powerful cycle of âwork-and-spend,â in which consumer norms accelerated dramatically. People needed to work more to purchase all the new products being churned out by a globalizing consumer economy. And they responded to their stressful working lives by participating in an orgy of consumer upscaling. There was an upsurge in luxury goods consumption, but now the aspiration to own these status items had become widely shared. Over the last thirty years, real consumption expenditures per person have doubled, from $11,171 to $22,152.
Conclusion
Recent trends in working hours are almost astonishing. Unlike the century between 1850 and 1950, when productivity improvements translated into considerable reductions in hours of work, the last three decades have witnessed steady increases in work time.
Between 1969 and 2000, the overall index of labor productivity per hour increased about 80 percent, from 65.5 to 116.6 (1992 = 100). That index represents economic progress, indicating that the average worker in 2000 could produce nearly twice as much as in 1969. Had we used that productivity dividend to reduce hours of work, the average American could be working only a little more than twenty hours a week. Thatâs the most extreme assumptionâall productivity increases channeled into shorter hours.
And what if that had happened? Our material standard of living would have stabilized. Americans would be eating out less, house size wouldnât have grown by 50 percent, and kitchen cabinets might still be made of formica. We also wouldnât be heating up the climate as rapidly, because expensive gas-guzzling SUVs wouldnât11 have become so popular. We wouldnât need to replace our computers every two to three years either, which might not be such a bad thing, at least from an environmental point of view. (A recent report suggests that the average computer uses a total quantity of material resources equivalent to the average car, or more.)
Itâs worth noting that stable incomes do not mean static consumer choice. Certainly, Americans would be consuming a different mix of goods and services than in 1960. But in the aggregate, taking all productivity growth as leisure time would have led to a stable real level of income.
But rather than focus on the stability of income, why not consider the temporal gains? The normal workweek could go as low as 20 hours, plus seven weeks of vacation. Two-income households with children could easily do without paid child care, because their work-time commitments would be low. People would have plenty of time for community and volunteer work, perhaps meaning less need for government social spending. It would be easy to pursue a passion, like playing music or woodworking, or quilting, or fishing.
We could become lifelong learners, or make up for our chronic national sleep deficit. All that free time could also go into pleasurable activities that provide additional income or consumptionâlike gardening, or making crafts for sale, or building furniture, or sewingâbut that increasingly few people have time for now. There would also be fewer work-related expenses which would make stable salaries more bearable.
Americans could actually get back to eating dinner together, talking, and visiting friendsâall activities that have been pushed out by excessive work time. From todayâs vantage point, a time-surplus society may seem utopian, almost unnatural. But thatâs only because weâve been going at 24/7 for too many years and have lost sight of other possibilities.
Itâs not too late to stop and smell the roses. The time has come to take back our time.
TABLE 2 / Growth in Annual Hours and Productivity in the United States, 1967â2000 (average annual change)
1Growth in Annual Hours and Productivity in the United States, 1967â2000. For a comprehensive discussion of this debate, see the work of J.B. Schor.