NBER WORKING PAPER SERIES
MEASURING TRENDS IN LEISURE: THE
ALLOCATION OF TIME OVER FIVE DECADES
Mark Aguiar
Erik Hurst
Working Paper 12082
http://www.nber.org/papers/w12082
NATIONAL BUREAU OF ECONOMIC RESEARCH
1050 Massachusetts Avenue
Cambridge, MA 02138
March 2006
We thank Susanto Basu, Gary Becker, Kathy Bradbury, Kerwin Charles, Raj Chetty, Steve Davis, Jordi Galí,
Rueben Gronau, Dan Hamermesh, Chad Jones, Ellen McGrattan, Bruce Mayer, Kevin Murphy, Derek Neal,
Valerie Ramey, Richard Rogerson, Frank Stafford, and seminar participants at the Minneapolis Federal
Reserve, the Cleveland Federal Reserve (NBER EFG/RSW meeting), the University of Rochester, the
University of Wisconsin’s Institute for Poverty Research Summer Institute, NBER Summer Institute in Labor
Studies, the University of California at San Diego, the University of California at Berkeley, the University
of Chicago, Columbia University, Boston College, Harvard, Wharton, and the University of Maryland. We
thank Dan Reichgott for research assistance. Hurst would like to acknowledge the financial support of the
University of Chicago’s Graduate School of Business. The views expressed in this paper are solely those
of the authors and do not reflect official positions of the Federal Reserve Bank of Boston or the Federal
Reserve System. The views expressed herein are those of the author(s) and do not necessarily reflect the
views of the National Bureau of Economic Research.
©2006 by Mark Aguiar and Erik Hurst. All rights reserved. Short sections of text, not to exceed two
paragraphs, may be quoted without explicit permission provided that full credit, including © notice, is given
to the source.
Measuring Trends in Leisure: The Allocation of Time Over Five Decades
Mark Aguiar and Erik Hurst
NBER Working Paper No. 12082
March 2006
JEL No. D12, D13, J22
ABSTRACT
In this paper, we use five decades of time-use surveys to document trends in the allocation of time.
We find that a dramatic increase in leisure time lies behind the relatively stable number of market
hours worked (per working-age adult) between 1965 and 2003. Specifically, we show that leisure
for men increased by 6-8 hours per week (driven by a decline in market work hours) and for women
by 4-8 hours per week (driven by a decline in home production work hours). This increase in leisure
corresponds to roughly an additional 5 to 10 weeks of vacation per year, assuming a 40-hour work
week. Alternatively, the "consumption equivalent" of the increase in leisure is valued at 8 to 9
percent of total 2003 U.S. consumption expenditures. We also find that leisure increased during the
last 40 years for a number of sub-samples of the population, with less-educated adults experiencing
the largest increases. Lastly, we document a growing "inequality" in leisure that is the mirror image
of the growing inequality of wages and expenditures, making welfare calculation based solely on the
latter series incomplete.
Mark Aguiar
Federal Reserve Bank of Boston
600 Atlantic Avenue
Boston, MA 02210
mark.aguiar@bos.frb.org
Erik Hurst
Graduate School of Business
University of Chicago
Hyde Park Center
Chicago, IL 60637
and NBER
erik.hurst@gsb.uchicago.edu
2
1. Introduction
In this paper, we document trends in the allocation of time over the last 40 years. In
particular, we focus our attention on measuring how leisure time has evolved within the United
States. In commonly used household surveys designed to measure labor market activity (such as
the Current Population Survey (CPS) and the Panel Study of Income Dynamics (PSID)), the only
category of time use that is consistently measured is market work hours.
1
As a result, leisure is
almost universally defined as time spent away from market work. However, as noted by Becker
(1965), households can also allocate time towards production outside the formal market sector.
To the extent that non-market (home) production is important and changing over time, leisure
time will be poorly proxied by time spent away from market work. By linking five decades of
detailed time-use surveys, we are able empirically to draw the distinction between leisure and the
complement of market work. In doing so, we document a set of facts about how home production
and leisure have evolved for men and women of differing work status, marital status, and
educational attainment during the last 40 years.
The main empirical finding in this paper is that leisure time—measured in a variety of
ways—has increased significantly in the United States between 1965 and 2003.
2
When
computing our measures of leisure, we separate out other uses of household time, including time
spent in market work, time spent in non-market (home) production, time spent obtaining human
capital, and time spent in heath care. Given that some categories of time use are easier to
categorize as leisure than others, we create four distinct measures of leisure. Our measures range
from the narrow, which includes activities designed to yield direct utility, such as entertainment,
socializing, active recreation, and general relaxation, to the broad, namely, time spent neither in
1
In some years, the PSID asks respondents to individually report the amount of time they spent on household chores
during a given week. These data are exploited by Roberts and Rupert (1995) to document a decline in total work,
which, for the overlapping periods, is consistent with the trends documented in this paper.
2
We provide a formal definition of leisure in Section 3.
3
market production nor in non-market production. While the magnitudes differ slightly, the
conclusions drawn are similar across each of the leisure measures.
Using our preferred definition of leisure, we find that leisure has increased by 7.9 hours
per week on average for men and by 6.0 hours for women between 1965 and 2003, controlling for
demographics. Interestingly, the decline in total work (the sum of total market work and total
non-market work) was nearly identical for the men and women (7.9 and 7.7 hours per week,
respectively). These increases in leisure are extremely large. In 1965, the average man spent 61
hours per week and the average women spent 54 hours per week in total market and non-market
work. The increase in weekly leisure we document between 1965 and 2003 represents 11 to 13
percent of the average total work week in 1965. Valuing time at 2003 market wages, the increase
in leisure has a market value of $5,000 to $5,500 per adult in annual terms. Aggregating over the
adult population, this represents 8 to 9 percent of total GDP in 2003. If we assume the after-tax
market wage represents the marginal rate of substitution between consumption and leisure, to a
first order approximation the increase in leisure is equivalent to 8 to 9 percent of 2003
consumption expenditures.
The adjustments that allow for greater leisure while satisfying the time budget constraint
differ between men and women. Men increased their leisure by allocating less time to the market
sector, whereas leisure time for women increased simultaneously with time spent in market labor.
This increased leisure for women was made possible by a decline in the time women allocated to
home production of roughly 11 hours per week between 1965 and 2003. This more than offset
women’s 5-hours-per-week increase in market labor.
3
We also analyze changes in leisure by educational attainment. We find that men and
women with more than a high school education and men and women with a high school education
or less all increased leisure time between 1965 and 2003. However, while the level of leisure in
3
The magnitudes we present in the introduction correspond to changes in time use conditional on demographic
changes, as shown in Figures 2–5.
4
1965 was roughly equal across educational status, the subsequent increase in leisure was greatest
among less-educated adults. Similarly, we document that the cross-sectional distribution of
leisure time has fanned out over the last 40 years. Given that the least-educated households
experienced the largest gains in leisure, this growing “inequality” in leisure is the mirror image of
the well-documented trends in income and expenditure inequality. The fact that the least-educated
experience the most leisure poses an empirical puzzle for the standard model that relies solely on
income and substitution effects: The time-series evidence suggests that rising incomes induce
greater leisure, while the recent cross-sections suggest that higher incomes are associated with
lower levels of leisure.
2. Related Literature
Three classic book-length references on the allocation of time are Ghez and Becker
(1975), Juster and Stafford (1985), and Robinson and Godbey (1999). The latter is most closely
related to our study. It uses the same time-use surveys we use from 1965, 1975, and 1985, as well
as some additional time-use information from the early 1990s.
4
Our paper adds to the earlier
results of Juster and Stafford and Robinson and Godbey by documenting the growing dispersion
in leisure as well as analyzing a longer time series. We also consider alternative leisure
aggregates. Several other studies have explored the trends in housework, including Bianchi et al.
(2000) and Roberts and Rupert (1995). In addition to extending the sample of Robinson and
Godbey through the late 1990s, the former work contains a nice summary of the existing
sociology literature on housework. The latter uses the market work and housework measures in
the PSID, as does Knowles (2005), who focuses on relative work hours (at home and in the
market) of spouses in younger households. For a popular but controversial study that draws
4
Juster and Stafford (1985) fully examined unconditional and conditional time use in the United States using the 1965
and 1975 time diaries. In the first edition of their book (1997), Robinson and Godbey extended the analysis of Juster
and Stafford by examining the trends in time use across 1965, 1975, and 1985. In their second edition, Robinson and
Godbey added a short chapter entitled “A 1990s Update: Trends Since 1985”. In that chapter, they briefly discuss how
unconditional measures of time in the early 1990s compare with unconditional measures of time use from earlier
decades. However, their discussion does not include the conditional time-use analysis that is done in this paper.
