Family labor supply, commuting time,
of the Tokyo Metropolitan Areaq
sity, Ki
R23
a m
ll-tim
n’s pa
are g
ated t
resen
e Tok
Several studies have investigated the relationship be-
tween women’s employment and housing purchases. Yos-
hikawa and Ohtake (1989), Fortin (1995), and Del Boca and
Lusardi (2003) empirically examine how the working
1051-1377/$ - see front matter � 2011 Elsevier Inc. All rights reserved.
be released due to the terms of usage of the data. I thank Kazuyasu
Sakamoto, Yasuhiro Sato, and seminar participants at Nagoya University
and Toyama University for helpful comments. This paper was presented
at the 2009 Spring Meeting of the Japanese Economic Association at Kyoto
University (June 6, 2009) and FESAMES 2009 Meeting at University of
Tokyo (August 3, 2009). Remaining errors are my own. This research is
supported by the Japan Society for Promotion of Science Grant-in-Aid for
Scientific Research (Grant Number B-19330053 and C-20530188).
⇑ Fax: +81 11 706 4947.
E-mail address: abey@econ.hokudai.ac.jp
1 Compton and Pollak (2007) also show that the concentration of power
couples in MSAs declined slightly between 1990 and 2000.
Journal of Housing Economics 20 (2011) 49–63
Contents lists available at ScienceDirect
Journal of Housing Economics
journal homepage: www.el
doi:10.1016/j.jhe.2010.12.001
Area in Japan. I present an optimizing model of family la-
bor supply and residential choices that incorporates an
important institutional feature of the Japanese labor mar-
ket: the decision by women to work full time or part time.
The Tokyo Metropolitan Area is also unique in that real
1940 to 1990. They attribute this change to the colocation
problem, i.e., the difficulty of finding a residence conve-
nient to both spouses’ workplaces. In contrast, Compton
and Pollak (2007) analyze panel data and show that power
couples are no more likely than other types to migrate to
large metropolitan areas. Accordingly, they argue that
colocation is not a major cause of the concentration of
power couples in large MSAs.1 Neither of these studies,
however, is very explicit about the relationship between res-
idential decisions and housing prices.
q This paper uses microdata of the Employment Status Survey (ESS)
made available through the Ministry of Internal Affairs and Communi-
cation of Japan under Article 33-2 of the Statistics Act. Microdata cannot
Keywords:
Commuting
Family labor supply
Part-time work
1. Introduction
How is women’s employment rel
cal distribution of households? The p
this question in a specific context: th
o the geographi-
t paper explores
yo Metropolitan
estate prices are much higher for locations close to the
Central Business District (CBD).
Costa and Kahn (2000) show that highly educated
‘‘power couples’’ in the United States became more likely
to locate in the Metropolitan Statistical Areas (MSAs) from
JEL classification:
J22
R21
gests that high commuting costs could be one of the main obstacles for women’s full-time
employment in places like the Tokyo Metropolitan Area.
� 2011 Elsevier Inc. All rights reserved.
Yukiko Abe ⇑
Graduate School of Economics and Business Administration, Hokkaido Univer
a r t i c l e i n f o
Article history:
Received 19 January 2010
Available online 6 January 2011
a b s t r a c t
In this paper, I build
incorporates the fu
explain why wome
cantly in areas that
and residential decisions: The case
ta 9 Nishi 7, Kita-ku, Sapporo 060-0809, Japan
odel of family labor supply and residential choices that explicitly
e or part-time work decisions of married women. The model can
rticipation patterns in full-time and part-time work vary signifi-
eographically close but differ in real estate prices. The model sug-
sevier .com/locate / jhec
factors make full-time and part-time work very distinct,
whereas the middle ground between these two options is
almost nonexistent. The model shows that the housing
market operates in a way to further strengthen this dis-
crete feature in the labor market.
Discreteness in the labor market can sometimes be mit-
igated by other ‘‘markets.’’ For instance, if annual full-time
earnings are higher than part-time earnings, in a multi-
period setting full-time workers retire early but part-time
workers retire late. Then lifetime earnings from the two
The remainder of the paper is organized as follows. Sec-
50 Y. Abe / Journal of Housing Economics 20 (2011) 49–63
status of married women is related to housing purchases
by their households. In the regional science literature,
Mok (2007) asks whether incomes of spouses are pooled
or not in deciding housing location. Using the Canadian
Census data for the Toronto Metropolitan Area, she finds
that incomes are not pooled for couples without children.
Iwata and Tamada (2008) build a model of commuting
behavior of married women and argue for a non-mono-
tonic relationship between the wage rate of working mar-
ried women and their commuting time. Sakanishi (2007,
2008) builds a model in which household location is deter-
mined by the husband’s choice and the wife chooses her
employment status given the husband’s location.
In this paper, I present a model of family labor supply
that permits married women to choose from full-time
and part-time job opportunities. The commuting costs,
wage levels, and working hours for the two employment
options are explicitly modeled so as to reflect the real con-
straints faced by couples in Japan. The model explains why
the distribution of women’s employment status differs sig-
nificantly across prefectures in the Tokyo Metropolitan
Area. Although the theoretical model is constructed to ex-
plain the facts for the Tokyo Metropolitan Area, the impli-
cations of the model are broadly consistent with empirical
patterns observed in the New York Metropolitan Area: evi-
dence from the New York area is reported in the Appendix.
The paper makes two novel contributions to the litera-
ture. First, it explicitly considers the optimization over
married women’s work (either full-time or part-time)
and residential location. In modeling this, the ‘‘discrete-
ness’’ of these two choices is particularly emphasized. This
element of the model is especially relevant in the Tokyo
Metropolitan Area, where full-time work is more common
among women living close to the CBD than among women
living in the suburbs. Such regional differences in partici-
pation are not observed for men (Section 2). By modeling
the family’s housing location and labor supply decisions
simultaneously, this model provides a natural explanation
for such patterns.
Second, the paper highlights the importance of com-
muting costs in understanding women’s work in Japan.
Several important measures (legal and otherwise) to pro-
mote women’s participation in the workplace have been
adopted in Japan in recent decades (e.g., the Equal Employ-
ment Opportunity Law, the Maternity Leave Law, and var-
ious policies to make balancing work and family easier).
Nonetheless, the proportion of women working in regular
full-time jobs in Japan has not increased much, especially
over age 40. Most of the increase in women’s employment
after middle age has taken the form of non-regular, part-
time work (Abe, in press). Commuting costs (including
the time and energy necessary for commuting) have not
fallen much for female workers in the past several decades,
which may explain the sluggish increase in regular full-
time employment for women.
The unique aspect of the model is that it explicitly
incorporates the discreteness of the labor market in terms
of full-time or part-time status. In Japan, full-time wages
are much higher than part-time wages. Furthermore, it is
difficult for women who left the labor force at childbearing
age to return to regular employment (Ueda, 2007). These
tion 2 presents stylized facts on regional differences in la-
bor market participation. Section 3 presents a model that
explains the stylized facts. Section 4 provides numerical
examples of the model and Section 5 concludes. The
Appendix reports evidence from the New York Metropoli-
tan Area.
2. Stylized facts
2.1. Discreteness in women’s employment choices: full-time
and part-time work in the Japanese labor market
To introduce the stylized facts to be explained in the
theoretical model, I begin with an overview of full-time
and part-time work in the Japanese labor market. In sub-
sections 2.1 and 2.2, participation in the two types of
wage–salary earners is considered: (1) regular full-time
employees and (2) part-time and casual employees.3
Part-time and casual employees comprise the typical non-
regular workforce in Japan and their numbers have in-
creased dramatically in recent decades. According to the
ESS, among women aged 25–59 years, the number of part-
time or casual workers was 4.5 million in 1987 and 7.4
2 Abe (2009), albeit in a different setting from this paper, presents a
model that explains this point.
3 In the data used in this paper, part-time and casual are the terms for
employment status in the workplace. Both part-time and casual employees
are non-regular workers. There is no clear distinction in tasks or working
conditions for the two types of workers, although casual workers are
generally younger than part-time workers. Students who work part time
normally work under casual status. In the questionnaire of the Employment
Status Survey (ESS, Statistics Bureau, Ministry of Internal Affairs and
Communications of Japan), casual workers are referred to as ‘‘arbeit
workers’’.
employment statuses are more equalized than for single-
period earnings, implying that intertemporal substitution
mitigates discreteness in a single period.2 What is unique
about the housing market in this paper is that it operates
to magnify the discreteness in the labor market.
The property whereby high-wage workers live close to
the CBD whereas low-wage workers live farther from the
CBD is common in many monocentric models because the
bid-rent gradient is higher for high-wage earners. By con-
trast, the model in this paper assumes that all men work
at the CBD while women have the option to engage in
part-time employment for which the commuting time is
zero. In this model, the distance from the CBD creates larger
differences on women’s behavior than for men’s. The large
discrepancy in women’s behavior is caused by the discrete-
ness in the labor market, magnified by the housing market.
.
3
.
4
8
Regular Part-time
arning
urce:
of the
in Jap
Y. Abe / Journal of Housing Economics 20 (2011) 49–63 51
million in 2007.4 Regular employees are those with no term
limits in their employment contracts. Part-time and casual
employees typically hold a contract of no longer than 1 year
in duration, although such contracts are often renewed
multiple times. Regular employees are eligible for employ-
er-provided fringe benefits, including public pension contri-
butions, a corporate pension, and employer-provided health
insurance; part-time and casual employees are much less
likely to obtain such coverage. Regular employees normally
work 40 h/week or more, whereas part-time and casual
employees work fewer hours. Hourly wage rates are higher
for regular employees than for part-time employees; accord-
ing to the statistics reported in Abe and Tanaka (2007), the
part-time/full-time wage ratio for female workers in 2001
was 51%.5 Hereafter, the term ‘‘part-time workers’’ denotes
both part-time and casual workers.
Although the part-time/full-time wage gap is one
source of the discreteness between the two options, a more
0
.
1
.
2
0 2 4 6
sh
ar
e
Annual e
Fig. 1. Earnings distributions of regular and part-time female employees. So
in the ESS, and the horizontal axis of the figurecorresponds to the midpoint
that appear in the figure are not evenly spaced. Data are from all regions
important source is probably the so-called 1.03 million yen
ceiling. The Japanese tax and Social Security system, as
well as the fringe benefit policies of employers, creates
an environment in which married women who are depen-
dents of their husbands have a strong incentive to restrict
their earnings to less than 1.03 million yen (for recent evi-
dence on this, see, e.g., Nagase and Nawata, 2005;
Akabayashi, 2006; Sakata and McKenzie, 2006; Abe, 2009).
As a result, the annual earnings of married women are
quite different depending on their employment status.
Fig. 1 shows the earnings distributions for full-time and
part-time working women from the ESS in 2007. Part-time
earnings are heavily concentrated around 1 million yen,
whereas full-time earnings show no such a concentration.
Earnings are higher for full-time than for part-time work-
ers; in the ESS data used in Fig. 1, the average earnings of
4 During the same period, the number of female regular workers in the
same age range changed from 7.5 to 8.9 million.
5 An extensive literature on part-time jobs in Japan provides evidence of
these issues; see, e.g., Houseman and Osawa (1998, 2003), Abe (2003),
Gaston and Kishi (2007).
female full-time employees are 3.3 million yen, whereas
those of female part-time employees are 1.0 million. Fur-
thermore, 93% of female part-time employees have earn-
ings of less than 1.5 million yen, whereas 56% of female
regular employees earn less than this amount. These are
the main features of the discreteness in the two employ-
ment choices for women in Japan.
2.2. Participation in full-time and part-time work by region
To understand the features of regional variations in wo-
men’s participation, I use the data from the ESS for 2007.
The ESS is a household survey and the region-based statis-
tics reported here are based on the place of residence (not
on place of work).6
I compare the participation patterns for the four prefec-
tures in the Tokyo Metropolitan Area: Saitama, Chiba, To-
kyo, and Kanagawa. Tokyo prefecture includes the Tokyo
0 2 4 6 8
s (in million)
EES 2007 (published version). Notes: The earnings are surveyed in intervals
earnings interval. Since the intervals are not evenly spaced,the midpoints
an.
CBD; the other three prefectures are generally suburban,
although they do contain some large cities such as Saitama
(in Saitama prefecture), Yokohama and Kawasaki (in
Kanagawa prefecture).
In the following discussion, three measures of participa-
tion are used: the employment-to-population ratio (E–P
ratio), the regular employment ratio, and the part-time
employment ratio. The regular employment ratio is the
number of regular employees divided by the population,
for which regular employees include regular wage and sal-
ary earners, as well as executives of private corporations.
The part-time employment ratio is the number of part-
time and casual employees divided by the population.
The three measures are plotted against age for married
women (Fig. 2A–C) and the regular employment ratio for
6 The ESS is conducted every 5 years by the Ministry of Internal Affairs
and Communications of Japan. In 2007, the survey was conducted for adults
in about 450,000 households; the size of the original sample was
approximately 1 million persons aged 15 and over. The ESS data do not
contain information on the location of the employment.
married men is plotted against age (Fig. 2D), using
microdata from the ESS.7 The data show several interesting
patterns. First, the E–P ratio for women below age 35 is high-
er in Tokyo than in the three other suburban prefectures.
This ratio is similar in all four prefectures after age 35. The
types of female employment observed in the four regions
are quite different: regular employment is more common
in Tokyo, whereas part-time employment is more common
in the other three prefectures, independent of age. Second,
over 80% of married men work as regular full-time employ-
ees, while less than 30% of married women work under such
a status (Figs. 2B and D). Finally, the regular employment ra-
0.
4
0.
5
0.
6
0.
7
0.
8
0.
9
30 40 50 60Age
Saitama Chiba
Tokyo Kanagawa
A. E-P Ratios of Married Women in Tokyo MA
0.
1
0.
2
0.
3
0.
4
30 40 50 60
Age
Saitama Chiba
Tokyo Kanagawa
B. Regular Employment Ratios of Married Women in Tokyo MA
0.
1
0.
2
0.
3
0.
4
30 40 50 60Age
Saitama Chiba
Tokyo Kanagawa
C. Part-time Employment Ratios of Married Women in Tokyo MA
0.
5
0.
6
0.
7
0.
8
0.
9
1
30 40 50 60Age
Saitama Chiba
Tokyo Kanagawa
D. Regular Employment Ratios of Married Men in Tokyo MA
Fig. 2. Employment-population ratio, regular employment ratio, and
part-time employment ratio of married women and men in four
prefectures in the Tokyo Metropolitan Area. Source: ESS 2007
(microdata).
52 Y. Abe / Journal of Housing Economics 20 (2011) 49–63
tio for men does not differ much across the four prefectures,
although it is slightly lower in Tokyo than elsewhere. Over-
all, the regional differences in women’s employment are
clearly greater.8 It is natural to expect that these significant
differences are caused by endogenous choice of residential
location. Housing prices are higher in Tokyo and full-time
employees earn much more than part-time employees. On
the other hand, work schedules are less flexible for full-time
jobs. High earnings by full-time working wives may be used
to finance the high cost of housing close to the CBD. House-
holds in which the wife does not work full time will locate
farther from the CBD to spend less on housing.
2.3. Educational composition and labor supply choices across
prefectures
Most large-scale datasets for Japanese workers do not
contain information on both wage and place of residence,
so it is difficult to directly test the hypothesis in the previ-
ous subsection. In this subsection, I therefore use educa-
tion as a proxy for wage levels. It is well known that
education is highly correlated with full-time wages. I first
look at the proportion of educated people among residents,
and then consider the proportion of educated workers
among residents.9
Fig. 3 shows the proportion of people with more than a
university education (16 years or more) among residents in
the four prefectures by age group. Among both men and
women of all age groups, the proportion of highly educated
persons (‘‘power’’ individuals in the terminology of Costa
and Kahn (2000) and Compton and Pollak (2007)) is high-
est in Tokyo, followed by Kanagawa, Chiba, and Saitama.
These data support the notion that high-wage earners tend
to live closer to the Tokyo CBD.
The educational distribution does not show whether la-
bor supply is related to residential choice because the popu-
7 I omit the E–P ratio and the part-time employment ratio for married
men because regional variations in the E–P ratio are smaller than those in
the regular employment ratio and because the part-time employment ratio
is low.
8 More generally, it is known that women’s participation in the labor
market, especially with respect to regular full-time work, has significant
regional variations within Japan (Takeishi, 2007; Abe et al., 2008). Regional
variations are small for men.
9 Costa and Kahn (2000) define ‘‘power couples’’ as couples in which
both spouses have more than a college degree. The available data do not
allow me to investigate the combination of educational attainment of
husband and wife. In this paper, ‘‘power’’ men and women are those who
have at least a university degree, and ‘‘power working’’ men and women are
those who have at least a university degree and work.
0.
1
0.
2
0.
3
0.
4
0.
5
30 40 50 60
Age
Saitama Chiba
Tokyo Kanagawa
Share of residents with a university degree or more
by Prefecture: Women
0.
5
0.
6
Share of residents with a university degree or more
by Prefecture: Men
ture in
Y. Abe / Journal of Housing Economics 20 (2011) 49–63 53
0.
2
0.
3
0.
4
30 40 50 60
Age
Saitama Chiba
Tokyo Kanagawa
Fig. 3. Share of residents with a university degree or more, by prefec
0.
2
0.
3
0.
4
0.
5
0.
6
0.
7
30 40 50 60
Proportion of power men amon
本文档为【Family labor supply, commuting time, and residential decisions】,请使用软件OFFICE或WPS软件打开。作品中的文字与图均可以修改和编辑,
图片更改请在作品中右键图片并更换,文字修改请直接点击文字进行修改,也可以新增和删除文档中的内容。
该文档来自用户分享,如有侵权行为请发邮件ishare@vip.sina.com联系网站客服,我们会及时删除。
[版权声明] 本站所有资料为用户分享产生,若发现您的权利被侵害,请联系客服邮件isharekefu@iask.cn,我们尽快处理。
本作品所展示的图片、画像、字体、音乐的版权可能需版权方额外授权,请谨慎使用。
网站提供的党政主题相关内容(国旗、国徽、党徽..)目的在于配合国家政策宣传,仅限个人学习分享使用,禁止用于任何广告和商用目的。