Paper Prepared for the Conference on Marriage in Cosmopolitan China, Hong

Kong University, July 4-6, 2011

(Re)emergence of
Late Marriage in New
Shanghai

Yong CAI and WANG Feng

6/20/2011
Please dont cite or distribute without consulting the authors for the latest version
and for permission

(Re)emergence of Late Marriage in New Shanghai

Yong CAI, University of North Carolina at Chapel Hill

caiyong@unc.edu
WANG Feng, The Brookings Institution and the University of California, Irvine
fwang@brookings.edu

Abstract

Over a time span of fifty years, young people in Shanghai, Chinas largest metropolis, produced two
phases of late marriage, once under the socialist rule in the late 1970s and another during the post
socialist era, in the late 1990s. These two late marriage regimes however were products of vastly
different forces and they inform in important ways of the drastic social changes in urban China in the
last half century. Using data from Chinas 2005 Population Sample Survey, this paper documents the
emergence and the reemergence of late marriage in Shanghai, and discusses the underlying forces
driving these two phases as well as their social and political implications.

(Re)emergence of Late Marriage in New Shanghai

1. Introduction: Marriage and Social Transformation in Urban China

In the last half century, two phases of late marriage emerged in urban Shanghai, Chinas largest and
most cosmopolitan urban center. Mean age at first marriage first rose from 20 for females and 22 for
males in the early 1950s to 25 for females and 27 for males by the late 1970s, the highest among any
Chinese cities. Yet, mean marriage age dropped precipitously after 1980, by two years both for females
and males, only to be followed by a gradual rise again, approaching the level of the previous peaks by
the early 21st century. In urban Shanghai, as elsewhere in China, a late marriage regime has reemerged.
The forces underlying these two late marriage periods however are vastly different. The first late
marriage era, while initially rooted in individual choices and preferences, was a product of forced
collective synchronization, under the socialist governments late marriage campaign as part of a forceful
birth control program. When such a requirement was withdrawn, as was the case in 1980, individuals
reacted and readjusted their behaviors, which led to a sudden and significant drop in marriage age
within a short time period. The second late marriage era, which emerged gradually over two decades
since the early 1980s and is continuing today, has brought mean marriage ages back to the level of the
previous peaks, especially for males. Unlike the first phase of late marriage, this reemergence of late
marriage is no longer the product of government policy enforcement, but due entirely to individual
volition and choices, conditioned by a post-socialist society that has commodified everything from
property to the human body. The reemergence of the late marriage regime in Shanghai, following a rise
and then a drop in marriage age during the previous decades, tells a vivid story of the social
transformation in China, seen from one of the most long-lasting and important social institutions.
In contrast to historical European societies where marriage and mate choice were based on individual
preferences (Schofield 1985, Macfarlane 1986), marriage until recently in Chinese history had long been
outside the confines of individual decision and preferences. Romantic stories featuring individual love
ending both in happiness but more often in tragedies are not few, but they mostly existed in novels and
plays as stories if not fantasies. In everyday life and for most people, marriage was not based on
romantic love and individual choice, but under the authority of the collective, for the purpose of
continuing the family lineage line and for building economic, social and political alliances (Yang, 1957;
Lee, Wang, and Ruan 2001). Marriages were arranged by the elders, either parents or other kin. While a
significant proportion of poor men remained bachelors throughout their lives, virtually no women were
unmarried by the age of 20 (Wolf and Huang 1980, Lee and Wang 1999). Early and universal female
marriage was therefore a defining feature of the Chinese society that had not only demographic
importance but also far reaching social and economic implications (Malthus 1798; Hajnal 1953, 1982;
Lee and Wang 1999).

Suppression of individual desires and decision in marriage formed a major source of social discontent
among the young, and Chinas marriage and family traditions were consequently among the main
targets of Chinese social revolutions in the last century. Chinese revolutionaries, especially the
communists, took reforming marriage practices as a high priority in their search for modernity. Footbinding, arranged marriage, gender inequality and many other old family institutions were deemed
obstacles of Chinas rejuvenation thus must be smashed. Womens emancipation, freedom of marriage,
and liberation from patriarchal control were among the important tenets of the May 4th movement in
1919. The Nationalist government, with its bourgeois support and western connection, promoted
womens new role through its New Life Movement, especially in urban parts of China from the mid1930s to the end of the 1940s. At the same time, more forceful social reforms that embraced gender
equality were carried out in the mountainous areas under communist control.
It was not until the founding of the Peoples Republic of China in 1949 that the traditional marriage
system and womens role inside the Chinese family were under a wholesale onslaught. A new marriage
law was among the first acts of the communist regime. The first (and only in the first 30 years) legalistic
action of the communist regime, the Marriage Law was enacted on April 13, 1950, only months after the
founding of the Peoples Republic. The law prohibited arranged and forced marriage, and promoted
gender equality. It abolished polygamy, concubinage, and little daughter-in-law marriage in which girls
were adopted by their future husbands families at a very early age and raised to be daughters-in-law.
The law stipulated that marriage should be a free decision between an adult male and an adult female.
According to the law, there should be no interference from any third person, and no one should seek
economic benefits from marriage. The law also provided venues and protections for divorce and
remarriage. However, even with lofty ideas and sometimes drastic actions, changes on the ground was
probably more gradual and evolutional than revolutionary. The law was extremely vague on punishment
on violation and noncompliance. In addition, although divorce was a protected legal option for marriage
exit (some even referred the Marriage Law as the divorce law), the social pressure and cultural prejudice
kept divorce at a very low rate until recently.1 The implementation of the law instead relied mostly on
social mobilization, mass campaign, education and persuasion.
In overthrowing the old marriage regime, the communist government took away the power of marriage
arrangement from the Chinese family but did not return that power fully to individuals (Croll 1981).
Whereas a decisively historical transformation did take place, for marriages to move from parental
decision to individual choice (Whyte 1990, 1993, Davis and Harrell 1993), the Chinese state kept an
1

There was an early spike of divorce in association with the new marriage law and mainly as a result of ending the
arranged marriages occurred earlier.

active role, both in encouraging individual choice for choosing a partner and later on, in dictating the
time of marriages. The social motivation of the new Marriage Law makes this state takeover fits well
with Foucaults (1998) biopower, in which the state exerts control over individuals under the name of
promoting better life. Moreover, the political structure of China and its socialist economic settings made
it easier for Chinese government to directly inject itself in peoples daily life.
In the 1970s, the wan (late marriage), xi (longer birth interval), shao (fewer births) campaign
aimed at slowing down Chinas rapid population growth was among the most blatant government
interventions in individual marriage of any modern societies. In a decade time, mean age at first
marriage rose by more than two years, creating the first phase of Chinas late marriage (Coale 1989,
Wang and Tuma 1993). Such a forceful policy interfering individual marriage not only ran counter to the
communists early promises of individual freedom but also invited fierce resistance and complaints. It
was therefore a glaring contradiction in demographic terms that while China stepped up its birth control
efforts in the late 1970s culminating in the announcement of the one child policy, it at the same time
also removed the requirement of late marriage age, both in the same year, 1980, resulting in a shortlived baby boom and literally cancelling out the initial birth control impact of the one child policy (Coale
1989, Feeney et al. 1989).
With government intervention leaving the scene, marriage choices have once again been returned to
individuals and their families. Over the two decades following the removal of the government stringent
requirement on marriage age in 1980, a new era of late marriage among Chinese young people ushered
in. Such a return is seen for China as a whole, but most prominent in urban China, in cities such as
Shanghai. Unlike the previous late marriage era, when the government was a major actor, the recent
late marriage phase reflects more of individual choices and preferences, and therefore informs us more
of the quiet yet profound social changes occurring in China.
In the remainder of this paper, we use urban Shanghai as an example to examine the reemergence of
the late marriage in contemporary China. We will first introduce the setting and the data we use in this
study, followed by two analytic sections, first an analysis and a comparison of the two phases of late
marriage, and second an analysis of the individual variations underlying the reemergence of the late
marriage, aimed at understanding how this phase of late marriage has taken place. We end the paper
with a brief discussion and conclusion.

2. Setting and Data Source

The rise of Shanghai epitomizes economic and social transformations in China over the last two
centuries. Shanghai was one of the original five ports opened to western commerce under the Treaty of
Nanjing of 1842. Those treaty ports quickly became offshoots of Western culture, and cradles of Chinas

modernization. At the beginning of the 20th century, Shanghai had grown to be the economic and
industrial center of China. The economic reform reopened Shanghai to the world, and Shanghai is now
one of the economic and financial centers of the world. Today, Shanghai is the largest city in China, the
most recent census counts its population over 23 million.
Shanghai is also the pioneer of Chinas social development. Chinas population control efforts, as well as
a push for late marriage were first implemented in Shanghai. Shanghai started advocating for population
control and late marriage in 1958 (Guo 1996). Shortly after the ending of the Great Leap Forward famine,
Shanghai put full force into population control. In December 1963, Shanghai announced its operational
definition of late marriage: for urban couples, around 25 for females and around 30 for males; 2 for
rural couples, 23 for females and around 25 for males.3 On surface, late marriage was only advocated as
advanced behavior. Those who followed governments call to delay their marriages by meeting the
late marriage age criterion were rewarded with financial and social welfare benefits. In reality, the late
marriage age criterion was often implemented as a compulsory rule.4 Those who did not answer to the
call were often denied their application for marriage registration. They also faced enormous pressure at
the workplace and within the residential committee. In fact, because city government had tight control
of peoples survival, such as job assignment and housing provision, compliance could be relatively easily
enforced. Shanghais approach of using late marriage in combination with longer birth interval and
fewer children were adopted as a national policy in the 1970s. Most provinces adopted similar but lower
age requirement for late marriage, most 23 for females and 25 for males5
The revision of the Marriage Law in 1980, which raised the legal minimum marriage ages by 2 years to
20 for females and 22 for males, and tightened up the interpretation of marriage age from traditional
year counting to full year counting6, signaled a major change in Chinas approaches in regulating
population and governing the society. Since the adoption of late marriage as a mean of population
control in Shanghai in the 1960s and at the national level in the 1970s, the de facto minimum marriage
ages, often specified in local regulations or decrees, were several years higher than the minimum legal
marriage ages stipulated in the Marriage Law of 1950. While there was nevertheless a drastic increase in
marriage age, the resistance to such a push was strong, and the compliance was far from perfect.
2

Age requirement for first birth was set at 26.

1978 8 25
23 27 25
4
1971
5
1974 12
25 23 4

6
Chinas traditional age counting starts with age 1 at birth, and increases one year at each Chinese New Year. By
the traditional calendar, when someone turns age 18, she or he could be only 17 or 16 in full year. On June 26,
1950, Chinese government published a memo titled Few Issues in the Implementation of the Marriage Law. In
responding to the question on how marriage age should be calculated, the answer states that age could be
calculated as full-year or could be calculated according to traditional counting.
3

Raising the de jure minimum marriage ages to lower than the de facto level was, on the one hand, a
continued recognition of the importance of late marriage in fertility control, but on the other hand, it
was also an acknowledgment of the challenges in implementing an unpopular policy that interfered with
one of the most basic decisions of individual lives. The revision therefore reflected a shift in the
governance philosophy of the Chinese government at the time, namely to return some freedom to the
population in the spirit of building a freer society based on voluntary participation, and to follow the rule
of law rather than administrative fiats (Banister 1987). Both changes in association with the new
marriage law were consistent with many other economic and social reform policies that helped
revitalize and rejuvenate a society that was still leaving behind the turmoil of the Cultural Revolution.
The revision of Marriage Law in 1980 did not mean that the Chinese government abandoned its efforts
to use late marriage as a method for population control. The enforcement of late marriage continued
and was incorporated and formalized in local birth control regulations. According to Shanghais Birth
Planning Regulations of 1981, late marriage ages were still set at 5 years above the legal minimums. In
1990, Shanghai further lowered its late marriage ages to 3 years above the legal minimums. According
to the 1990 revision of Shanghais Birth Planning Regulation, couples meet late marriage age criteria
could have one week of vacation time for their honeymoon and more time for paternal leaves.7
While there were no major changes in its regulations on marriage since the publication of Birth Planning
Regulations in 1990, Shanghai has experienced drastic changes in almost every aspects of individual life.
Shanghai is now leading China in the level of economic prosperity. In 1990, Shanghais GDP per capita
was just over $1,000; by 2010, it had reached $10,000. It is also leading China in demographic
transformation. Shanghai now has Chinas lowest fertility and mortality, with a total fertility lower than
1.0 and life expectancy higher than 80 years. Yet, the population size of Shanghai has expanded from 13
million to 23 million in just two decades. The vast majority of the population increase was due to a flood
of migrants from other parts of China. By the early twenty-first century, over a quarter of Shanghais
population were migrants from elsewhere in China, every two out of five workers are migrants, a third
of all marriages involved a non-native partner. Among babies delivered in Shanghai in 2007, over 40
percent were from parents who were not native Shanghainese (Ruan 2009). Shanghai in other words is
now a new Shanghai.

1981 8 27
25 25 23 25
23 15
1984 10
1987
4
2000
25 23 24 1990
3
25 23 24
3

To understand changes in marriage in Shanghai today, we use data from Chinas Inter-census Survey of
2005 collected by the National Bureau of Statistics of China (Zhang et al. 2005). In China, a full
population census takes place every ten years at years ending with a 0, and an inter-census population
survey, also referred as 1% population sample survey or mini-census, takes place at years ending with a
5. A mini-census is very similar to a full census except for its probabilistic design. The mini-census uses
Probability Proportionate to Estimated Size (PPES) sampling method to ensure representation at the
provincial level, with resident groups in rural area and neighborhood committees in urban area as
primary sampling units (called enumeration districts). Target population are all current residents in the
selected enumeration districts, with or without a local household registration, and include those who
maintain local household registration but were away at the time of enumeration. The survey
questionnaire is very similar to the census questionnaires of 2000 and 2010 with 35 items at the
individual level, and 20 items at the household level. The data used in this analysis is a 15% sub-sample
of the mini-census for Shanghai.
We use data from the 2005 mini-census to examine changes in marriage age since 1950. The survey not
only contains information on the current marital status, but also year and month of the first marriage.
Computer algorithm is used to match marriage partners within a household according to their
relationship with the household head, for generating data needed for spousal information. Because
marriage might have a selective effect on or through mortality, for example, people in urban areas tend
to marry later and live longer than people in rural areas, age at marriage for early years calculated from
the survey might present some bias.8 However, data examination indicates such a bias is minimal in
Shanghai and does not change the trend of marriage changes as we shall describe below.9

3. From Collective Synchronization and Individual Liberalization

Two phases of late marriage age are clearly evident among the population in Shanghai. In Figure 1, we
present changes of male and female ages at first marriage in two separate panels, one for males and
one for females. Four summary statistics are selected to present both the central tendency and
variability in changes of marriage age at first marriage. Age at first marriage has been on the rise since
1950, with trends basically parallel for males and females, and females mean age at first marriage about
2 years younger than males. The revision of Marriage Law in 1980 divides the rise of age at first
marriage in Shanghai from 1950 to 2005 into three stages: pre-1980, 1980-1990, and post-1990.
8

If marriage has a negative effect on mortality, a retrospective survey would over-estimate average marriage age
for older generations if mortality is not controlled for, and vice versa.
9
At the national level, comparing to female mean age of marriage estimated from the 1988 Fertility Survey,
estimates based on the 2005 mini-census at national level is about 0.5-1.0 year higher for women married
between 1940 and 1970, but the effect is just the opposite, but smaller for men. The exact reasons for such a gap
and for gender difference require further investigation, but one main suspect is the large rural/urban gap in
mortality and marriage in China. No major difference is found for Shanghai, for both male and female.

Figure 1. The Rise of First Age at Marriage, Shanghai 1950-200510

During the first phase of rising marriage age, between 1950 and 1980, the trend was almost linear:
average age at first marriage rose for about 5 years in these three decades. Unlike what has been
observed previously at the national level that there was a sudden acceleration in the increase of mean
age at marriage in the 1970s (e.g. Coale 1989; Wand and Yang 1996), as a direct result of the later,
longer, fewer policy, the increase in marriage age in Shanghai appears to be steady and continuous
with only minor interruptions (e.g. the Great Leap Forward). The overall increase was similar for males
and females, but the pace was faster in 1950s for males. Between 1950 and 1960, mean age at first
marriage rose almost 3 years for males, from 21.9 to 24.8. At the same time, the mean age at first
marriage for females only rose by 2 years, from 20.3 to 22.3. However, in the 1960s and 1970s, the
increase of mean age at first marriage was faster for females than males, reflecting in part the drive
behind this push for late marriage, which was to reduce fertility. The average age at first marriage
reached 26.2 in 1970 and then 27.0 in 1980 for males, and 23.3 in 1970 and 25.1 in 1980 for females.
The swift rise in at first marriage was reversed by the revision of Marriage Law in 1980. The reversal was
just as speedy as the rise: by 1990, mean age at first marriage had dropped to 25.3 for male and 23.3 for
female, nearly two years each for males and females respectively.
10

Source: Calculated using data from the 2005 1% Population Change Survey.

With no obvious policy connection, mean ages at first marriage resumed its rise in the early 1990s,
presumably a result of socioeconomic change and cultural change. From 1990 to 2005, it increased from
25.3 to 26.5 for male and 23.3 to 24.1 for female, close to the level reached in the late 1970s and
producing the second phase of late marriage age among individuals in Shanghai.
The trends depicted by median ages at first marriage for both sexes match closely with those of mean
ages at first marriage, with the exception of a divergence after the decline in the 1980s. Age at marriage
typically has a positively skewed distribution with more people marrying at older ages than younger
ages, thus a higher mean value than the median. The late marriage push that imposed forced
synchronization on marriage age in Shanghai made marriage age distributions more symmetric than
normally expected. Only when government loosened its control of marriage age, did we see the
reemergence of a positively skewed marriage age distribution, especially on the male side.
The more interesting story of marriage age change in Shanghai is not in the rise-fall-rise of the mean
ages, but in how such trends are formed by individual behaviors, which are seen in the measures of
variations. One set of such variation measures are the 95 and 5 percentiles, meaning ages at which the
latest five percent (95 percentile) and the earliest five percent (5 percentile) of individuals married. The
95 percentile trends in Figure 2 are parallel to the mean and median trends for the most part, until the
mid-1980s, when the liberalization of marriage age control took place. For males, the 95 percentile
continued its rise, only to stabilize around 35 years around 2000; for females, the liberalization spurred a
small decline, but then the trend reversed and settled at around 30 years. Unlike the mean and median
trends, whose values from recent years are approaching but have yet to reach the historical highs of
early 1980s, the top percentile, represented by very late marriages, has surpassed its historical highs. In
other words, more people began to marry at ages substantially higher than what is suggested by the
means or medians.
Similarly, the 5 percentile trends in Figure 2 are parallel to the mean and median trends, but tell the
story on the low end of marriage age. It suggests that underage marriage was indeed widespread in
China. Even in Shanghai, where its social and cultural surrounding makes its population more susceptive
to the idea of late marriage and where the government has a more effective control over its population,
there were still a non-miniscule proportion of people marrying before reaching legally minimum age.11
However, the relaxation of marriage age control in 1980s had a smaller effect on the 5 percentile than
on both the mean and median. The stabilization of the 5 percentile after 1990 suggests that there is
likely a culturally defined low end of age at marriage.

11

Because this study is based on a retrospective survey, we cannot separate marriages happened in Shanghai and
marriages took place somewhere (before people moved to Shanghai).

Figure 2. Mean and Standard Deviation of Age at First Marriage, Shanghai 1950-200512
One main difference between the recent and previous rise of marriage age is that the recent rise
happened at the same time when the age range of marriage also expanded. This second phase of late
marriage age is thus in clear contrast to the earlier one, when marriage age range stayed the same or
even shrunk. Such a contrast indicates that the rise of marriage age between 1950 and 1980 was a result
of forceful government policy that effectively synchronized peoples marriage age, but the recent one
was a product of increased individualization and liberalization, ushering in an era of late marriage that is
based more on individual volition and choice.
The shift from collective synchronization to individual differentiation is seen most clearly with another
measure of variability, namely the standard deviation of ages at first marriage. As shown in Figure 3,
parallel to the rise of marriage age up until the late 1970s is a sharp decline in the variability of marriage
age (right figure). The variability declined from 3.5 years to below 2.5 years in around 1980, before a
reversal starting in 1980, most likely due to the loosening of marriage age control and the start of
individual liberalization in marriage timing. By the late 1990s, the standard deviation has returned to the
pre-late marriage era. With the Marriage Law set the minimum marriage ages, the source of variability
in age at first marriage over time is in the upper end of marriage age, i.e. what are the proportions of
12

Source: Calculated using data from the 2005 1% Population Change Survey. Dotted/Dashed are fitted medianspline trend lines.

10

people marring at old ages. In the 1960s, there was an increase of proportion of women marring in the
late 20s, but the proportion declined in the 1970s. Starting in the 1980s, the upper end of age at first
marriage for females settled at about 30, and for males about 35 (as shown in the right panel Figure 2).
The launch of the late marriage policy first greatly increased the variability of age at first marriage in
early 1960s, especially for females. The unnatural increase in mean ages was then associated with a
decline in variability. The liberalization of marriage age control has reestablished the traditional age
variability by 1990, and since then, marriage age for males in Shanghai has shown further variability, a
result of a greater degree of individual differentiation.

Figure 3. Age Difference within First Married Couple by Marriage Year, Shanghai 1950-200513
Similar collective synchronization followed by individual liberalization can also be observed in the age
difference within a couple (Figure 3). While the new Marriage Law of 1950 broke many traditions of the
Chinese marriage system, it nevertheless created a new type of female age hypergamy, by stipulating a
two-year age difference between male and female minimum age at marriage. The institutionalization of
age hypergamy and the push for late age made it extremely difficult for a female to marry a younger
husband because she would have to wait several extra years for her groom to reach legal/de facto
minimum age at marriage. Such an effect can be seen in Figure 3: the proportion of female marrying a
13

Source: Calculated using data from the 2005 1% Population Change Survey. Only first married couples (both
sides) are included in the calculation.

11

younger husband plunged in the 1950s, and then stabilized at about 20%14. It can be argued that this is a
result of institutionalized hypergamy.
The forced synchronization of marriage age up to the 1980s not only stabilized a two-year norm of the
age difference between males and females, it also drastically reduced the variation among couples in
their age difference. As shown in the right panel of Figure 3, by the late 1970s, the variability measure
of the age difference between spouses reached its lowest point. Concurrent with the emergence of the
late marriage age in the 1990s, age difference between spouses also increased drastically. In contrast to
the high variability during the period leading to the first late marriage era, prior to 1970, when the
variability in spouse age difference is driven in part by older wives than husbands (as shown in the left
panel of Figure 3), the latest rise in spousal age difference is driven by more older men marrying
younger wives, an important feature of the recent individualization.
The rise of age at first marriage in Shanghai since 1950 shows both the cultural resilience and cultural
change in the Chinese marriage regime. Dramatic, sometimes chaotic social, economic, and political
changes in China since the founding of the Peoples Republic have had direct impacts on Chinas
marriage system. The changes in the marriage system however seem to be more an evolutionary than
revolutionary process. Even the forceful push of late marriage by government fiat only had a
temporary effect. Looking from a long-term perspective, age at first marriage in urban Shanghai was on
a course of steady and gradual rise. The underlying forces driving such a long-term trend, as we
document above, varied by historical time and institutional conditions.

4. Changing Patterns of Individual Behavior

Having delineated the two phases of late marriage in urban Shanghai and identified that a defining
feature of the reemergence is increasing individualization, our next questions are 1) what the effects of
such delayed marriage have on the marriage institution itself, namely whether the delay is also
associated with staying out of the marriage institution all at once, and 2) how such an increased
individualization plays out, namely who are the ones marrying the latest, if at all.
The connection between late marriage and the rising marriage age is obvious, but the delay may not
affect the other defining feature (along with early marriage) of the traditional Chinese marriage system,
namely universality of marriage. Even with delay, everyone still gets married, just at a later age. This is
largely true if we look at the current marital status by age in Shanghai (Table 1). In 2005, by age 35-39,
only 6 percent of males and 2 percent of females stay never married. By 40-44, virtually all women are
married, while only about 5% of men remain unmarried. The relatively high proportion of never married
men is also due to sexual imbalance rather than individual choice.

14

Trend analysis suggests that there is a modest trend of increasing from 1960s and on.

12

Table 1. Current Marital Status by Age, Shanghai 200515

Percentage
Sex
Male

Age Group

Never
Married

First
Marriage

Remarried

Divorced

Widowed

25-29
30-34
35-39
40-44
Total

3,437
3,393
3,258
3,490
13,578

38.6
12.0
6.1
5.4
15.6

60.8
85.6
89.8
87.9
80.9

0.2
0.7
1.9
2.8
1.4

0.4
1.7
2.1
3.8
2.0

0.0
0.1
0.1
0.2
0.1

Female 25-29
30-34
35-39
40-44
Total

3,597
3,328
3,071
3,262
13,258

23.6
5.7
2.0
0.8
8.5

75.3
91.0
91.6
91.1
86.9

0.5
1.5
3.3
3.7
2.2

0.6
1.7
2.8
3.7
2.2

0.1
0.1
0.3
0.8
0.3

Under the general picture of universal marriage, however, there is a social gradation of marriage in
relation to other socioeconomic indicators. Figure 4 presents the connection between education and
marital status. For age group 25-34, the proportion of never married rises gradually with educational
attainment. For age 35-44, the association between education and proportion of never married persists
though with a much higher proportion of ever married, suggesting at least some highly educated
women may choose to forgo marriage all together. Another interesting phenomenon is that divorce and
remarriage proportions are higher for those with a middle school or high school education.

15

Source: Calculated using data from the 2005 1% Population Change Survey.

13

80
60
40
20
0

20

40

60

80

100

Age 35-44

100

Age 25-34

Primary School Middle School

High School

Never Married

College

University

First Marriage

Primary School Middle School

Remarried

High School

College

Divorced

University

Widowed

Figure 4. Marital Status by Age and Education, Shanghai 200516

The change from a forced collective synchronization of marriage timing in the 1970s to the reemergence
of individual choices can be seen in the comparison of cumulative risk of marriage between the decades
of 1971-1980 and of 1996-2005 (Figure 5). In 1971-1980, the main difference is between those with
lowest level of education as one group and all others as another group. The differential effect of
education is shown primary as a delaying effect. Individuals with only a primary school education
married at much younger age, and at a much faster pace. But by age 35, individuals with higher levels of
education made up their lost time and reached a level of marriage similar to those with only a primary
school education. The effect is very similar for males and females.
In the decade of 1996-2005, the differential effect of education on marriage had become much more
evident and widespread, with marriage timing clearly differentiated by the level of educational
attainment. Large gaps are seen between those with lower than middle school education and those with
high school education, and between those with high school education and those with college degrees
and above. There is also another important change that suggests the weakening of universal marriage
for females: unlike what shown in the decade of 1971-1980, when females regardless of educational
16

Source: Calculated using data from the 2005 1% Population Change Survey. The Primary School category
includes those with no education. The University category includes those with professional and graduate school
degrees.

14

attainment eventually converge to universal marriage, the trend shown in the 1996-2005 decade reveals
that, for women with the highest levels of education, almost 7 percent of them would remain single at
the age of 45 if the risk of marriage stay at the level of 1996-2005. The differential effect of education on
marriage is more than a mere delay, either long delay led to missed opportunities, or some highly
educated women made a deliberate choice to stay single.

Figure 5. Kaplan-Meier Estimates of Cumulative Risk of Getting Married

Based on Empirical Data of 1971-1980 and 1996-2005, Shanghai17

The male side of the story for 1996-2005 decade is equally interesting. Given an imbalanced sex ratio
both historically and again in recent times (Coale and Banister 1994, Cai and Lavely 2007), and the
strong emphasis in Chinese culture on continuing the patriarchal family lineage, many have argued that
marriage is a symbol of social success and status (Lee and Wang 1998, Wang and Tuma 1996). There are
however several interesting crossovers in marriage risks among men with different education levels.
While individuals with a primary school education start to marry at earlier ages and at a faster pace than
other groups, they are surpassed by those with a middle school education at around age 28. In fact, for
17

Source: Calculated using data from the 2005 1% Population Change Survey. The Primary School category
includes those with no education. The University category includes those with professional and graduate school
degrees.

15

those with the lowest level of education, if they did not get married before age 30, their chance of
marriage is significantly diminished. At the same time, people with higher levels education, though with
a late start, move quickly to catch up. By age 35, the cumulative risk of getting married across three level
of educationmiddle school, college, and university are very similar to each other. However, there are
more late marriages among those with university level education, they would eventually move above
everyone else to have the highest proportion of ever marriage by age 45.
The divergence in marriage pattern between males and females among those with the highest levels of
education (college and university degrees) confirms our early observation that the reemergence of late
marriage in Shanghai is a result of individual choice within a strong cultural context of hypergamy. For
women, it is their choice of choosing not to marry, or at least not to compromise to make up their
missed opportunities in their 20s and 30s; for men, their status advantage is vividly in display as their
opportunity for marriage stays at a high level all the way to their 40s.

5. Conclusion
In a long time span of half of a century, the institution of marriage, a traditional social pillar of the
Chinese society, has witnessed profound changes. Behind such changes are both the force of the state
and social changes as embodied in changing family relationships and individual values and choices. In
Shanghai, and likely in China as a whole, 50 years of state engineered social development has greatly
weakened the traditional preference of early marriage. The first socialist marriage law of the 1950
liberated individuals from the rules of their family elders and led to a gradual increase in age at marriage,
but did not weaken the marriage institution itself. States enforcement of late marriage in the 1970s
pushed age at marriage to a historical high, erasing individual differences and creating a collective
synchronization. When such a strong intervention was withdrawn, marriage age first fell, followed by a
gradual and still ongoing rise. By the end of this half-century, a second height of marriage age emerged,
as seen in urban Shanghai. Unlike the previous peak of age at marriage, the recent one is featured by
increased individual choice and differentiation. And unlike the previous late marriage era, which
retrenched quickly when the state intervention was withdrawn, the current one carries its own
momentum and has no end in sight.
What has also emerged in association with the reemergence of late marriage in recent years is a move
away from the traditional model of universal marriage. Such a trend resembles what has taken place in
societies elsewhere, including those in East Asia, and departs from the patterns seen only two decades
ago in urban China, where marriage age delay was associated with a stalled wall, namely eventually
almost all young people still got married (Whyte 1993). At least for the highly educated women in
Shanghai, almost 7 percent will remain single at age 45 if the marriage trend observed in the decade of
1996-2005 continues. With the rapid expansion of Chinas high education, especially womens education,
if this trend of non-marriage continues, China will soon face a marriage revolution just like what has
happened in many developed countries.

16

The reemergence of late marriage in Shanghai takes place in the broad context of rapid socioeconomic
changes, among them liberalization, marketization, and globalization. Romantic relationship and selffulfillment have replaced family obligation and continuing lineage as the dominant tenets behind the
marriage binding young men and women. At the same time, life pressure in a highly competitive
environment that puts a price tag on virtually everything has turned mate choice and marriage decision
a rational calculation weighing costs and benefits. Although we were not able to perform a full-length
analysis of the connection between socioeconomic factors and marriage decisions, our analysis
demonstrates that there is a very strong relationship between education and marriage. Education has
direct effects on age trajectories of marriage likelihood; it also influences the relative position of men
and women in the marriage market.
Rising age at marriage is just one aspect of multi-facets change in Chinas marriage institution. Rising
pre-marital sex, cohabitation, divorce, extra-marital sex all pose challenges to a society that for
thousand years functioned on a relatively stable patrilineal, patrilocal and patriarchal family-marriage
system. Further analysis should not only look into marriage as outcomes, but also as agents of social
change, as changes in marriage have profound effects on many other aspects of the Chinese society.

17

References
Banister, Judith. 1987. Chinas Changing Population. Stanford, Calif.: Stanford University Press,.
Cai Yong and William Lavely. 2007. Child Sex Ratios and Their Regional Variation. In Zhongwei Zhao
and Fei Guo (eds.), Transition and ChallengeChinas Population at the Beginning of the 21st
Century, Oxford University Press. 108-123.
Coale, Ansley J. 1989. Marriage and childbearing in China since 1940. Social Forces 67(June): 833-850.
Coale, Ansley J. and Judith Banister. 1994 Five decades of missing females in China. Demography 31.3:
459 - 479.
Croll, Elizabeth. 1981. The Politics of Marriage in Contemporary China. Cambridge: Cambridge University
Press.
Davis, Deborah and Stevan Harrell. Eds. 1993. Chinese Families in the Post-Mao Era. Berkeley:
University of California Press.
Feeney, Griffith, Wang Feng, Zhou Mingkun, and Xiao Baoyu. 1989. Recent fertility dynamics in China:
Results from the 1987 One Percent Population Survey. Population and Development Review 15
(2): 297-322.
Foucault, Michel. 1978. The History of Sexuality. 1st ed. New York: Pantheon Books.
Guo Shenyang. 1996. Determinants of fertility decline in Shanghai: Development or policy? In China:
the Many Facets of Demographic Change (Alice Goldstein and Wang Feng, editors) Boulder:
Westview Press 1996. 81-96.
Hanjal, John. 1953. Age at marriage and proportions marrying. Population Studies 7: 111-136.
Hanjal, John. 1965. European marriage patterns in perspective. In D.V. Glass and D. E. Eversley
(eds.) Population in History: essays in historical demography. London.
Hanjal, John. 1982. Two kinds of preindustrial household formation system. Population and
Development Review 8, no. 3 (September): 449-494.
Lee, James and Wang Feng. 1999. One Quarter of Humanity: Malthusian Mythology and Chinese
Realities, 1700-2000. Cambridge, MA: Harvard University Press.
Lee, James, Wang Feng, and Danching Ruan. 2001. Nuptiality among the Qing nobility: 1640-1900. In
Asian Population History, Liu et al (eds.). Oxford: Oxford University Press. 353-373.
Macfarlane, Alan. 1986. Marriage and Love in England, Modes of Reproduction 1300-1840. Oxford:
Basil Blackwell.
Schofield, Roger. 1985. English marriage patterns revisited. Journal of Family History 10: 2-20.
Wang Feng and Nancy Tuma. 1993. Changes in Chinese marriage patterns during the twentieth century.
Proceedings of the IUSSP International Population Conference, Montreal, pp. 337-352.
Wang Feng and Yang Quanhe. 1996. Age at Marriage and the First Birth Interval: The Emerging Change
in Sexual Behavior Among Young Couples in China. Population Development Review 22.2 (June):
299-320.
Whyte, Martin King. 1990. Changes in mate choice in Chengdu. in Chinese Society on the Eve of
Tiananmen, eds. Deborah Davis and Ezra F. Vogel. Cambridge: Harvard University Press.
Whyte, Martin King. 1993. `Wedding behavior and family strategies in Chengdu. In Deborah Davis And
Stevan Harrell, eds. Chinese Families in the Post-Mao Era. Berkeley: University of California
Press.
Wolf, Arthur and Chieh-shan Huang. 1980. Marriage and Adoption in China, 1845-1945. Stanford:
Stanford University Press.

18

Yang, C. K. 1959. Chinese Communist Society: The Family and the Village. Cambridge: Harvard University
Press.
Zhang, Weimin, Li Xiru, and Cui Hongyan. 2005. China's Inter-census Survey in 2005. Paper presented
at the 22nd Population Census Conference, Seattle, Washington, USA, 7-9 March 2005.
http://www.ancsdaap.org/cencon2005/Papers/China/China.Zhang.Weimin.etal.pdf.

19