print logo

The Long-Hours Luxury

Monday, August 4, 2014

One factor that is often overlooked in the debate over causes of income inequality is a shift in the distribution of working hours. The rich now work more than the poor.

The United States and most other industrialized countries have experienced a rapid increase in inequality since around the 1970s. Unsurprisingly, the discussion of inequality has become politicized. Some deny the increase in inequality entirely, in spite of overwhelming evidence. Others exaggerate the magnitude of inequality, for example by claiming that all income gains have gone to the rich.

The estimates of income distribution by the Congressional Budget Office avoids several common methodological problems. The CBO adjusts for declining household size and uses a broader definition of income that includes pension and health care contributions by employers. Between 1979 and 2010 in the United States, the top 1 percent’s share of pre-tax income increased from an already high 9 percent to 15 percent. It is an overstatement that the middle class did not witness any gains, but it is certainly true that gains were slower than in the past. Real median income increased by a disappointing 36 percent between 1979 and 2010. During the same period, real earnings of the 1 percent of highest earners grew by an astonishing 280 percent. One-quarter of income growth in the United States from 1979 to 2010 flowed to the top 1 percent alone.

This rise of income inequality is intensely debated, but its causes are not well understood. Potential explanations include skill-biased technological change, globalization, tax cuts, the expansion of the financial industry, low-skill immigration, and measurement problems that exaggerate inequality. Many of these explanations have merit, and it is unlikely that one factor alone explains the phenomenon. It has, however, proven difficult to determine the relative importance of the various theories, with multiple causes competing to explain a single phenomenon.

Hours worked: One factor in income inequality

One factor that is often overlooked is a shift in the distribution of working hours. Historically, the poor and lower skilled tended to work more than the rich. This is no longer true. Today high-income earners and the highly educated tend to work more hours. This gradual shift is nowhere large enough to single-handedly account for the rise of income inequality, however, working hours are part of the explanation and have the advantage of being readily quantifiable. Changes in working time should not necessarily be viewed as an independent explanation for inequality; its decline is related to more fundamental causes of inequality, such as the relative decline in demand for low-skilled labor.

Most of the change is driven by changes in hours worked per employee, not by changes in employment rates.

The shift in the distribution of work and leisure has been documented by several time-use studies in the United States and Europe that ask participants to record their use of time in a diary during a reference period, such as a typical week. The most comprehensive survey of time use is the American Time Use Survey (ATUS), an annual survey collected by the Bureau of Labor Statistics since 2003.

An influential study by economists Mark Aguiar and Erik Hurst1 compared older time-use surveys with more recent data from 2003. The distribution of hours worked only changed slightly between 1965 and 1985. Between 1985 and 2003, however, a sizable divergence was observed, with hours worked declining among the lower skilled and increasing among the highly skilled. Among men with no high school degree, average work time declined from 44 to 33 hours per week. Among women with no high school degree, average work time declined from 18 to 15 hours per week. (Keep in mind that these averages include those who don’t work at all). By contrast, hours worked increased from 42 to 45 hours per week among college-educated men. Among college-educated women, average work time similarly increased 26 to 31 hours per week. In 1985, college-educated men as a group worked 5 percent fewer hours per week than men with no high school degree. By 2003, college-educated men worked almost 40 percent more.

Since the day still only has 24 hours, time use between work, leisure, domestic duties, and rest is a zero-sum game. Since the 1960s, time spent on household work in the United States has fallen sharply, especially among women. Technologies such as dishwashers and services such as home delivery of meals or groceries has reduced the need for household work. The decline in work was therefore accompanied by an increase in leisure among those with both high and low levels of education. Increased leisure among the less educated should, however, not necessarily be interpreted as an increase in well-being. Studies such as Sevilla et al (2012)2 also look at the quality of leisure. One method to estimate the quality of leisure is simply asking respondents to self-report how much they enjoy various activities. The low skilled have largely replaced work with cheaper leisure activities with below average levels of enjoyment, most importantly watching more television. In my view, it would be more accurate to write that an involuntary decline in working time has driven low-income groups into increased idleness rather than increased leisure. 

Economists Peter Kuhn and Fernando Lozano study3 the shift in work inequality using the Census Bureau’s Current Population Survey between 1979 and 2006. These estimates of hours worked are not as detailed as time diaries, but have a larger sample size. The results are similar to Aguiar and Hurst. Kuhn and Lozano divide the population into five equally sized groups based on income. They investigated the share of working-age male employees who work long hours, defined as more than 50 hours per week. For simplicity, let us compare the top 20 percent of the income distribution with the bottom 20 percent.

Between 1979 and 2006, the share of low-wage earners who worked long hours declined from 22 percent to 13 percent. In the same time period the share of high-wage earners who worked long hours increased from 15 to 27 percent. Results were similar when education rather than income is used to segment the labor market. Most of the change is driven by changes in hours worked per employee, not by changes in employment rates. For men lacking high-school education, one-third of the decline in hours is driven by reduced employment rates, while the rest is driven by decline in hours among the employed. Among college-educated men, the entire increase in the long hours is driven by those with employment working more hours.

Kuhn and Lozano write:

Perhaps the most striking feature … is the reversal in the cross-sectional relationship between hourly wages and work hours since the early 1980s: in 1979, the worst-paid 20 percent of workers were more likely to put in long work hours than the top 20 percent; by 2006, the top 20 percent were twice as likely to work long hours than the bottom 20 percent.

The reversal in working time is a major and little discussed shift in the labor market. The historical view of inequality is one where the poor masses toiled in the field or factories, while rich landowners or rentiers had the luxury not to work. Today we face a very different situation. Increasing numbers of the poor and lower skilled workers are excluded from the labor market.

The long-hours premium

In addition, putting in extra work hours tends to increase earnings exponentially rather than linearly. The work of skilled professionals and managers is often “nondivisible.” Nondivisibility of labor refers to situations where one employee working full-time is more productive than two separate workers doing the same task working half-time each. Life quality aside, one top manager who works 80 hours per week is preferable to the firm than hiring two managers who work 40 hours each. Workers who put in longer hours also signal their ambition and are more likely to beat the competition and receive promotions. Interestingly, the long-hours premium in the United States has almost doubled since the 1980s. Kuhn and Lozano write, “Overall, an extra hour beyond 40 was associated with a 1.2 percent increase in earnings in 1983–85 and with about a 2 percent increase by 2000–2002.”

Potential explanations for the rise of income inequality include skill-biased technological change, globalization, tax cuts, low-skill immigration, and measurement problems that exaggerate inequality.

It is not clear why this shift has occurred. The type of jobs in which people usually work long hours, such as finance, law, and medicine, have experienced income gains, which indirectly drags up the long-hours premium. Skill-biased technological change is probably another important explanation for a rising long-hours premium. In the past, the economy mostly produced goods. Increasingly, we have witnessed a shift in which the most important sectors of the economy produce knowledge. Rather than training new workers, there is an advantage in having experts, who have already invested thousands of hours to master their trade, work a few extra hours.

Nonpecuniary or nonmonetary motivations for work may also be important. Most people do not merely work for money. We work because we enjoy it, for social reasons, to learn, or simply to keep ourselves busy. This is particularly true in the United States, a country with a historically strong work ethic. According to a recent Gallup poll, 68 percent of Americans report that they would continue to work even if they won millions of dollars in the lottery.

Rewarding jobs

From the point of view of maximizing state revenue or GDP per capita, it doesn’t matter much if low income earners are pushed out of the labor force. But the decline of work is a human tragedy.

In simple economic models, working less and having more leisure increases well-being. A common but mistaken view of this reversal in work inequality is that it has benefited the low skilled because they can consume as much as before without having to work as hard. This ignores the complexity of human psychology.

Humanist theories of happiness, starting with Aristotle, have long argued that the key to life satisfaction is living a purpose-driven life and aiming for higher goals. Modern psychology similarly emphasizes work and purpose for a full life. Abraham Maslow viewed fulfilling one’s potential or “self-actualization” as the pinnacle level of happiness. Mihaly Csikszentmihalyi argued that people are happiest when they are in a state of “flow,” or a complete absorption in a challenging and intrinsically motivated activity.

As Larry Summers argued recently, the major economic challenge ahead will no longer be creating goods, but creating rewarding jobs for everyone. Recent trends in the United States and most European countries have not been encouraging in this regard. According to the Bureau of Labor Statistics, the employment-to-population ratio of those in their prime working age crashed from around 80 percent prior to financial crisis to around 75 percent in 2009. Today the figure is 76.7 percent, which means that only one-third of the lost ground has been recovered.

A low income is the economy’s way of signaling that certain work is not very valuable for society. The United States and Europe have a surplus of low skilled labor, which in most sectors is easily replaced by technology and capital. Today a few highly skilled factory workers using software and robots are producing more output than plants employing thousands not many years long ago.

I will not pretend to have any easy solutions to these fundamental problems. We may be facing an entirely new problem that falls outside the traditional debates of the previous century. The first technological shifts from an agricultural economy were effortlessly solved by market forces, to the surprise of most contemporary observers. Freed from the need of feeding the population, farmers switched jobs and moved to manufacturing and services, where their labor was demanded. We should, however, be careful about assuming some sort of technological determinism from one historic episode. We are moving into uncharted territory, with no guarantee that all workers will keep pace with technology forever.

Tino Sanandaji is a researcher at the Institute of Industrial Economics and holds a PhD in public policy from the University of Chicago.

FURTHER READING: Sanandaji also writes "Why Keynesianism Works Better in Theory Than in Practice." Gary S. Becker and Kevin M. Murphy consider "The Upside of Income Inequality." Thomas A. Hemphill, Waheeda Lillevik, and Mark J. Perry discuss "Confronting the U.S. Advanced Manufacturing Skills Gap." Steve Conover adds "The Myth of Middle-Class Stagnation." 



1. Aguiar, Mark, and Erik Hurst. "Measuring Trends in Leisure: The Allocation of Time Over Five Decades." The    Quarterly Journal of Economics 122, no. 3 (2007): 969-1006.

2. Sevilla, Almudena, Jose I. Gimenez-Nadal, and Jonathan Gershuny. "Leisure inequality in the United States: 1965–2003." Demography 49, no. 3 (2012): 939-964.

3. Kuhn, Peter, and Fernando Lozano. "The Expanding Workweek? Understanding Trends in Long Work Hours among  US Men, 1979-2006." Journal of Labor Economics 26, no. 2 (2008): 311-343.



Most Viewed Articles

3-D Printing: Challenges and Opportunities By Michael M. Rosen 10/19/2014
With physical copying now approaching digital copying in terms of ease, cost, and convenience, how ...
Government Sponsors Truthy Study of Twitter By Babette Boliek 10/21/2014
The debate over the National Science Foundation study of Twitter is getting off track. The sole issue ...
Why Privilege Nonprofits? By Arnold Kling 10/17/2014
People on the right view nonprofits as a civil-society bulwark against big government. People on ...
Chinese Check: Forging New Identities in Hong Kong and Taiwan By Michael Mazza 10/14/2014
In both Hong Kong and Taiwan, residents are identifying less and less as Chinese, a trend that ...
How Green Is Europe? By Vaclav Smil 09/30/2014
A superficial look might indicate great achievements. Yet a closer view reveals how far European ...