I recently noticed an item on Ezra Klein's blog about the "The Underworked American" which had been posted by Kathy G, I had to find out more. I am forever hearing about the overworked American, so I had to follow up on this contrary opinion. How could it be that we are underworked after all our griping about being overworked? I contacted Kathy G, snarfed a copy of the paper and set myself to the task of understanding.
The paper is based on time diary studies, has a pretty big sample size, and is demographically adjusted. All of these are good things. The authors, Mark Aguiar and Erik Hurst, seem to be pretty careful in the handling of the data and cautious in their analysis. They slice their datasets this way and that, and despite this, they get pretty much the same results. That is a good sign that the numbers are speaking for themselves. No matter how they slice things, leisure time in general has been rising since the base year of 1965, but it has been rising more for some than others.
The studies break up time into market work, non-market work, child care and leisure. Market work is work that one gets paid for. Non-market work is all that stuff around the house that just needs to be done or the place will be a falling apart pigsty. (Sometimes it is anyway, but that's another story). Child care is treated separately since it straddles the line between work and leisure. Then, there is everything else, goofing off, watching television, sleeping, eating Cheetosh, crocheting, model ship building, beer pong and so on. On a weekly basis, everything adds up close to 168 hours. That's the number of hours in the week, and that is the one thing that has not changed at all since 1965.
The result I was most interested in was the change in core market work, the work that gets paid for. According to the paper, core market work has stayed in the range of from 27.74 to 29.93 hours a week from 1965 to 2003. The low was in 1985, and the highs in 1965 and 1993. That is an extremely small range considering the number of "revolution in the workplace" articles that have been published in the same period. The chart below shows the data for both men and women and clearly shows a general decrease in core market work, and in other market work which includes commuting and lunch breaks. Core non-market work, or housework, is also decreasing, but the time spent on child care is rising.
If we look at the next chart, which shows the situation for men, we see more of what we might expect. The hours men work have been falling, with most of the drop between 1965 and 1985. The hours men spend on housework has risen slightly, from a slip to a sliver.
As the chart below shows, the number of hours women spent in market work has been rising, while time spent on house work has been falling. None of this surprising. We already knew this.
Looking at the numbers, we see the hours men worked falling from 42.09 in 1965 to 36.86 in 1985, a drop of more than five hours a week, over 10%. This was followed by a rise to 38.52 in 1993, still more than three hours less than in 1965 and a more recent fall to 35.54 in 2003. The real change was between 1965 and 1985. The work week for men fell from over 40 hours to well under 40 hours, even as Americans scoffed at the European movement towards a 35 hour work week.
Women, in contrast, started working longer hours, but this started later. Women worked between 18.83 and 19.84 hours from 1965 to 1985, but this jumped to around 22.5 hours by 1993, an increase of more than 10%. This new work week stayed there. Women added four hours to their work week shortly after men dropped five hours from their work week. Since there are more men in the work force than women, the total number of hours per working age person has decreased. The increase of women's hours in the work force have not compensated for the decrease of men's hours, so work time is fallling, and leisure time is rising. That should have been obvious.
In order to get some sense of what had happened in the latter third of the 20th century, I went to an alternate source of information, the Bureau of Labor Statistics. They had a handy dandy time series for hours in the work week as recorded by the official government of the United States, presumably based on asking employers and possibly employees about their hours. What we see in the chart below is a general fall in the hours worked in a week from 1965 into the 1980s. This chart is much closer to the men's hours chart above than the chart for women, but it is quite consistent in the trend towards shorter work weeks.
Once I started downloading and graphing series from the Bureau, it sort of took on a life of its own. There were two other series of interest. I was now curious as to how much employers were paying for each of those hours, how much that was worth in 2007 dollars and how much employees were earning in a work week. So, down through the internet flowed the Bureau's data on the Consumer Price Index, the CPI for urban regions, and down flowed the hourly wages. Into my spreadsheet they flowed, and out popped the chart below:
Here we see that the real value of an hour's worth of work, or the cost of an hour's worth of work if we are looking at it from the employer's point of view, was rising through the late 1960s, and it peaked in the annus mirabilis of 1973. The Vietnam War was ended, OPEC had raised oil prices, Nixon had resigned, and the real value of an hourly wage had peaked. The hourly wage fell, then leveled. Then came 1979, the annus ridiculous, with Three Mile Island, the second OPEC oil squeeze, and the Skylab fiasco. The hourly wage fell again, and it stayed down until Clinton's second term, and then it rose only slightly.
This was not great news if you were paid hourly, but it was possible to keep body and soul together on from $15.50 to $17.10 an hour (in 2007 dollars). Of course, this did not take into account the falling number of hours in the work week. If we look at the chart below, we can see the full impact. Yes, 1973 was the high point, and 1979 the turning point, but the 2007 value of the weekly wage had dropped from perhaps $650 to $570. That's more than a 10% cut in pay.
The end of the 1970s was the end of an era. The 1980s saw men's wages permanently lowered, more and more women entering the work force and children spending more time in day care. This rather bland chart explains much of why men grasped at social conservatism, seeking more respect now that they were being paid less, and fantasy economics, to justify their societal deprecation. The rise of day care led to all those bizarre child abuse cases involving clowns killing small animals and underground dungeons. There was a lot of anxiety around.
Since the 1980s, the value of an hour's work has been stagnant, and there are fewer and fewer hours of work available, but not everyone is paid by the hour. It is time to dive into a bit of accounting esoterica. There are two types of workers in the work force, exempt workers and non-exempt workers. If I remember correctly, non-exempt workers are paid by the hour, and they are not exempt from various overtime laws. In contrast, exempt workers are paid an annual salary, usually on a monthly or bi-weekly basis, and they receive the same money whether they work 20 hours or 80 hours a week. Exempt workers are generally better paid than non-exempt workers, and they get better bennies, but their jobs usually require higher levels of skill and better credentials.
If you consider matters from an employer's point of view, it makes perfect sense to minimize the number of hours when those hours have a fixed marginal cost, but to maximize the number of hours when those hours have no marginal cost. In an economically rational world we would expect to see increasing automation and optimization of jobs performed by hourly workers combined with increasing pressure on salaried, exempt, employees to maintain or even increase the number of hours they work.
If we go back to the study, we see that this is indeed the case. The authors looked at the market work, non-market work and leisure time for the different groups with different levels of education. They found quite a divergence between 1965 and 2003. While male college graduates cut their work week by 4.41 hours, male high school drop outs cut their work week by 18.02 hours. The chart below shows that between 1965 and 1985, the work week was shrinking for all men, but between 1985 and 2003, college graduates started working more while high school drop outs worked much less.
The situation is a bit different for women. Female college graduates added 4.26 hours to their work week. If a couple married after college graduation, the bride added just about the same number of hours to her work week as the groom subtracted. In the chart for women below, we see relatively little change between 1965 and 1985, except for women who had some college, but did not graduate. Starting in 1985, women college graduates started working more, even as women high school drop outs worked less. Women who attended college, but did not get a degree increased their work week over the entire period.
It would be fascinating to have more demographic information on this cohort. Was it the baby boom? Any explanation would involve sex, class, money and work, sort of like a Judith Kranz or Barbara Taylor Bradford novel.
So, what has happened to the overworked American? In fact, Americans are working less, but the Americans who write books, do studies, and generate newspaper columns and blog entries are working more. There is no time crunch for high school drop outs, but there is a money crunch as we know that high school drop outs earn much less than college graduates. I was pleased to see that the authors have provided some perspective for their study. In their conclusion they note:
"Costa [2000] documents that low-wage workers reduced their market work hours relative to high-wage workers between the 1890s and 1991. In particular, at the turn of the twentieth century, low-wage workers worked longer hours than high-wage workers. This differential disappeared by the early 1970s, and during the last thirty years high-wage workers supplied relatively more market hours."
The 20th century was the story of the declining work week. Farm production was mechanized and factory production was automated, minimizing the hours of labor required even as the cost of those hours rose. There was an opposite trend for jobs which were not measured by the hour, so exempt workers find themselves working longer weeks.
In their Middletown books, published in the 1920s and 1930s, the Lynds considered the business class and the working class with their different lives, attitudes, paychecks and work hours. These classes still exist, though they are less evocatively named non-exempt and exempt. While the business class, the exempt, is experiencing a time crunch, the working class, the non-exempt, is becoming the new leisure class, with shorter hours, high fructose corn syrup and reality television.
If you have waded this far through this commentary there is a good chance that you are in the 16+ demographic, and are longing for more leisure time. Even if you are not in this category, you likely have leisure time fantasies, perhaps revolving around tropical beaches or European cafes. Surely you are not wasting your precious leisure time reading the likes of this essay. But what is the new leisure class doing with its newly found hours?
The paper addresses this as well, but the answer it gives is rather sad. The chart below, copied from the original paper, is telling. From 1965 to 1993, leisure time, type two, rises. (The paper addresses four different types of leisure if you must know). Our unfortunate answer closely tracks the line of leisure. There it is, marked "tv". Just about every new hour of leisure gained has been spent watching television, at least until the rise of video games and the internet, post 1993.
Reading, as one might expect has been falling since 1965, but worse, socializing collapsed between 1965 and 1975 and has yet to recover. The chart title is most appropriate, "Breakdown of Leisure Activity".
It is impossible to argue with Mr. Aguiar and Mr. Hurst. Americans are working less and have more leisure time than ever. Time use surveys and the Bureau of Labor Statistics in its online wisdom bear this out. On the other hand, there is a growing chasm in leisure time between those who bill by the hour and those who sell their labor in bulk, the exempt and the non-exempt. Agricultural and manufacturing jobs are vanishing. The service jobs that remain lie in two categories, the low end hourly and the high end salaried. Jobs in the low end are under assault by automated tellers, computerized check out, household machinery and business reorganization. As we were told in the 1960s, a great big beautiful tomorrow of leisure awaits us, or at least most of us.
