How do womens education and career affect their decision on marriage and motherhood a case study for vietnam

ABSTRACT Nowadays, time to marry has been earlier among women in developing countries, and time of entry into first marriage has each particular effect on health issue of women and their

VIETNAM – NETHERLANDS PROGRAM FOR M.A IN DEVELOPMENT ECONOMICS

HOW DO WOMEN’S EDUCATION AND CAREER AFFECT THEIR DECISION ON

MARRIAGE AND MOTHERHOOD?

A CASE STUDY FOR VIETNAM

BY TRUONG UYEN PHUONG

Academic Supervisor

Dr TRUONG DANG THUY DECEMBER 15 th , 2016

ACKNOWLEDGEMENT

The first thing I would like to express is my deepest gratitude to my academic supervisor Dr Truong Dang Thuy Thanks for your enthusiastic, patient and dedicated support for guiding me to implement this thesis and overcome many difficulties during the entire process Once again, I am really appreciated your valuable encouragement

I would like to give my appreciation to all lecturers in Vietnam – Netherlands Program, who have provided me such useful knowledge that I can applied to this thesis as well as in my future’s job

I am grateful to all staffs of Vietnam – Netherlands Program and all of my friends for your help I deeply treasure all the moments we have and share with each other

Yet I have still had a lot of mistakes, I really expect all teachers to give sincere remarks to me to be better And the last one is our best wishes to all of you

ABSTRACT

Nowadays, time to marry has been earlier among women in developing countries, and time of entry into first marriage has each particular effect on health issue of women and their children However, there are not many researches focusing on this area, especially for the case of Southeast Asia countries Therefore, this paper is designated for filling this gap, and I choose the case of Vietnam for analysis to investigate the association between social factors such as education, ethnic, religion, and income and

on women’s first marriage and childbirth decisions The dataset is established from a random online survey with 505 respondents including men and women, but the purpose of this paper is not suitable for men, then male’s respondents are automatically excluded from the dataset Consequently, my sample consists of 304 women aged 18 to 66, which is divided into birth cohort in order to find whether there

is any difference in marriage or rearing children among generations Survival model analysis is applied

to give the probability of getting marriage at a specific time of women’s life The study found that educational level, income, time of first intercourse, promotion achievement have significant impacts on women’s marriage decision and fertility

CONTENTS

CHAPTER 1: INTRODUCTION

1.1 RESEARCH PROBLEM 7

1.2 RESEARCH OBJECTIVES 9

CHAPTER 2: LITERATURE REVIEW 2.1 THEORETICAL LITERATURE 11

2.1.1 Theory of Marriage & The division of Labor 11

2.1.2 Theory of Marriage Market 16

2.1.3 Theory of Fertility 18

2.2 REVIEW OF EMPIRICAL STUDIES 21

CHAPTER 3: DATA AND METHODOLOGY 3.1 DATA 25

3.2 METHODOLOGY 25

3.3 VARIABLES’S DEFINITION 27

CHAPTER 4: EMPIRICAL RESULT 4.1 EMPIRICAL RESULT 40

4.1.1.Statistics 40

4.2 RESULTS 71

4.2.1.Result for First-Marriage 71

4.2.2.Result for First-Birth 74

CHAPTER 5: CONCLUSION 5.1 MAIN FINDINGS & RECOMMENDATION 78

5.2 LIMITATION AND FURTHER STUDIES 80

REFERENCE 82

APPENDIX 84

LIST OF TABLES

Table 3.1 First-Marriage Variables Description (Education and Career Variables) Table 3.2 First-Marriage Variables Description (Social Background Variables) Table 3.3 First-Birth Variables Description (Education and Career Variables) Table 3.4 First-Birth Variables Description (Social Background Variables)

Table 3.5 First-Birth Variables Description (Social Background Variables)

Table 4.1 Summary statistics of First marriage

Table 4.2 Summary statistics of Women’s educational level at first marriage Table 4.3 Summary statistics of Women’s promotion achievement at first marriage Table 4.4 Summary statistics of Job movement at first marriage

Table 4.5 Summary statistics of Job at first marriage

Table 4.6 Summary statistics of Birth cohort

Table 4.7 Summary statistics of Residence at first marriage and first birth

Table 4.8 Summary statistics of being a chief income earner at first marriage Table 4.9 Summary statistics of acestor worship at first marriage and first birth Table 4.10 Summary statistics of Religion at first marriage and first birth

Table 4.11 Summary statistics of Birth order at first marriage and first birth

Table 4.12 Summary statistics of Father’s education at first marriage and first birth Table 4.13 Summary statistics of Mother’s education at first marriage and first birth Table 4.14 Summary statistics of Father’s job at first marriage and first birth

Table 4.15 Summary statistics of Mother’s job at first marriage and first birth Table 4.16 Summary statistics of First birth

Table 4.17 Summary statistics of Educational level at first birth

Table 4.18 Summary statistics of Promotion achievement at First birth

Table 4.19 Summary statistics of Job movement at First birth

Table 4.20 Summary statistics of Job at First birth

Table 4.21 Summary statistics of Living arrangement at First birth

Table 4.22 Summary statistics of Housework regularity at First birth

Table 4.23 Summary statistics of Abortion at First birth

Table 4.24 Summary statistics of Contraceptive knowledge at First birth

Table 4.25 Summary statistics of Age at First intercourse

Table 4.26 Summary statistics of Contraceptive type at First birth

Table 4.27 Summary statistics of Women’s wealth at first marriage and first birth Table 4.28 Results for First marriage by Exponential and Cox regression Model Table 4.29 Results for First birth by Exponential and Cox regression Model

(Educational and Career opportunities variables) Table 4.30 Results for First marriage by Exponential and Cox regression Model

(Social background variables) Graph 1 Probability of remaining single by women’s age

Graph 2 Probability of remaining single by women’s birth cohort

Graph 3 Probability of remaining single by women’s educational level

Graph 4 Probability of remaining single by women’s job

Graph 5 Probability of remaining single by women’s promotion achievement Graph 6 Probability of remaining single by women’s job movement

Graph 7 Probability of not having first child by Age

Graph 8 Probability of not having first child by birth cohort

Graph 9 Probability of not having first child by women’s education

Graph 10 Probability of not having first child by women’s job

Graph 11 Probability of not having first child by women’s promotion achievement Graph 12 Probability of not having first child by women’s job movement

Graph 13 Probability of marriage by educational level

Graph 14 Probability of marriage by promotion achievement

Graph 15 Probability of marriage by Job movement

Graph 16 Probability of marriage by Job

Graph 17 Probability of fertility by educational level

Graph 18 Probability of fertility by promotion achievement

Graph 19 Probability of fertility by job movement

Graph 20 Probability of fertility by Job

This trend in research analysis is properly comprehensible with visible evidence of postwar era in numerous countries The more investment in education and opportunities for employment has induced women to look for work in labor market, which leads to releasing from financial support of their husbands and great success in career

Moreover, marriage age has increased significantly in many developed countries around the world According to U.S Census Bureau, 2010, the median marriage age for women and men in 1950-1960

was 20 and 23 respectively; it increased to 27 and 29 in 2013 The situation is more serious for the case

of Germany, Netherlands, Denmark, United Kingdom, South Korea, Taiwan and many other countries with the range from 29 to 32 A great number of studies had been done for Western Germany, for instance, the paper of Diekmann (1990) found that the expansion of education has move the median age of marry upward almost one year

Although various researches have been conducted to identify the main factors of this trend among advanced countries, (e.g Blossfeld & Huinink, 1991; Cherlin, 1980; Diekmann, 1989; Elder & Rockwell, 1976; Hoem, 1985; Hoem & Hoem, 1987; Hogan, 1978; Huinink, 1987; Marini, 1985), very few studies are undertaken for less developed countries

Finally, marriage and fertility are connected processes, assuming that fertility often takes place within marriage and contraceptive practices are non-existent, there is an inverse relationship between time at first marriage and fertility Specially, in a Southeast Asian country like Vietnam, where traditional value is highly evaluated, woman must be married before having her first baby As educational level affects family formation, it also has direct impact on pregnancy decision Women’s increased control over fertility, a better chance to access to higher education, and a fall in discrimination will offer women a stable income, Goldin and Katz (2002), Blau and Kahn (1997, 2000), so they tend to earn more and more to prepare for children’s life

In Viet Nam, the age of first marriage of women is 22.8 years of age This figure did not change in the last one and a half decade (source), while that of men increased from 25.2 to 26.2 from 1999 to 2009 Age of first marriage of Vietnamese women is comparable to Southeast Asian countries, for example Cambodia Thailand 21, Malaysia 25.7, Indonesia 22.3, but quite low comparing to that of developed countries, for example Canada 29.1, UK 30 and the Netherlands 30.4 Women in these developed countries obviously have higher education and career opportunities compared to Vietnamese women One question arises is that whether Vietnamese women delay their optimal time for first marriage and birth change when having more education and career opportunities And what is the association

between education and delayed first marriage and childbirth? The answers to this question is quite important for a range of policies, including family planning, schooling and education services, and the planning of health care and child care services

This paper is also to find out the answer for the questions and help to diminish the anxiety and distress women have to encounter when their time of first children is delayed more than they or society expects

1.2 RESEARCH OBJECTIVES

This study focuses on the question of how the improvement in women’s education affects their marriage behavior and child-bearing decision, which is built on the “New home economics” theory A highly positive relation between years of schooling and marriage age is broadly accept; in other words, higher educational level is a causal factor for marriage postponing We will take into account some hypothesis such as “independence hypothesis”, “specialization hypothesis”, “human capital effect”,

“institution effect” in order to answer the basic question is whether higher education level delays marriage or if it also reduces marriage intensity In order to answer this question, I use the technique of

“Survival analysis” to examine how the probability of entry into first marriage and motherhood changes over ages

Particularly in this study, the application of survival analysis will help investigating how education and career change the rate of women entering first marriage and motherhood This is to provide information

on the potential benefits of policies that improve women education and career opportunities

In this paper, I intend to conduct an online survey for collecting data since there are no available sources on the needed information for this estimation, particularly the data on age at first marriage or first birth of women Based on economic theory on marriage or childbearing and previous empirical study, I identified several variables, including age-independence, social class, level of education, and participation in the educational system and cohort membership in the analyses These variables are asked directly in the questionnaire and expected respondents are women aged from 18 or higher

This study is designed into 5 main parts to theoretically and empirically analyze how women education

or career development as well as social background affect to their decision on marriage and childrearing

Chapter 2 postulates the social framework of theory that this paper is relied on Particularly, the most noticeable theory is the “New home economics” of Gary Becker (1981), which shed light on the determinants of marriage and demand for children

Chapter 3 represents the pattern through which data on women is collected and variables definition as well as model estimation to contribute a reliable implication on this area

Chapter 4 provides comprehensive results for the relationship that we have supposed from the beginning until this part It is expected to be consistent with the available theory

Chapter 5 gives a final conclusion based on the transparent results in part 4, from this perception; this

paper will commit its limitation and infer some further studies

2.1.1 Theory of Marriage & The division of Labor

In this section, I present the determinants of the benefits of marriage compared to single life for one man and one woman This will be the basis for analyzing the optimal timing of marriage

Consider two persons, a male (M) and a female (F) to observe whether they should marry each other or stay alone “Marriage” is assumed to be the action of M and F share the same household and according

to Gary Becker (1981), the incentive for both men and women to marry is the gaining of marriage life;

in the aspect of economics, they can increase their utility after getting marriage Utility depends not only on the purchased goods and services in market place, but also on the commodities produced by each household These commodities include the quality of food, the quality and quantity of children, reputation, entertainment, friendship, love and health status Importantly, these goods are not marketable or transferable between households, but only among members in the same household Consequently, they cannot be measured as a usual manner of other output, but we assume a single aggregate (Z) is a combination of all these commodities Maximizing utility thus becomes equal and similar for each person to maximize the amount of Z that he or she receives The production function of each household which connects its total output of Z to different inputs is displayed below:

Z = f (𝑥1,…., 𝑥𝑚; 𝑡1, … , 𝑡𝑘; 𝐸) (1)

In which 𝑥𝑖 are various market goods and services, 𝑡𝑗is time inputs of different household members, and E are “environmental” variables The budget constrain for the 𝑥𝑖.can be written as:

∑ 𝑝𝑚 𝑖𝑥𝑖 = ∑ 𝑤𝑘 𝑗𝑙𝑗+ 𝑣 (2) Where 𝑤𝑗the wage rate of the jth member is, 𝑙𝑗 represents how much time a man spends on working in the market sector, and v is the property income 𝑙𝑗 and 𝑡𝑗are related by the basic time constraint:

𝑙𝑗 + 𝑡𝑗 = T (3) Where the total time of each member is denoted by T, substituting (3) into (2), a single full income constraint can be constructed by the combination of the goods and time constraints as (4)

∑ 𝑝𝑚 𝑖𝑥𝑖+ ∑ 𝑤𝑘 𝑗𝑡𝑗 = ∑ 𝑤𝑘 𝑗𝑇 + 𝑣 = 𝑆 (4)

In which if the 𝑤𝑗 is unchanged, full income is appreviated by S – the maximum achievable income, there is an assumption that a decrease in household’s total output (Z) cannot make any members better off but some worse off Hence, to maximize the total output Z, each member would not hesitate to contribute his time and goods to this allocation And necessary conditions to maximize Z include:

𝑀𝑃 𝑥𝑖

𝑀𝑃𝑡 𝑗=𝑝𝑖

𝑤 𝑗 for all 𝑥𝑖 > 0 𝑎𝑛𝑑 0 < 𝑡𝑗 < 𝑇 (7) Thus, there must be an appropriate allocation and cooperation in time between the market and nonmarket sectors among each member If a man and a woman are married, their household is assumed

to contain only the two time inputs of them; in other words, we have ignored the time of children and other people living in the same household

From the equation (5) and (7), we can infer that female would specialize in nonmarket sector if 𝑤𝑚/𝑤𝑓

or 𝑀𝑃𝑡𝑓/𝑀𝑃𝑡𝑚are sufficiently large

Similarly, a single household allocates only his or her time between the nonmarket and market sectors

to satisfy equation (7) Specifically, single woman is more likely to work more than married woman because they do not have time and goods supplied by the others partner

If 𝑍𝑚0 𝑎𝑛𝑑 𝑍0𝑓represent the maximum outputs of single man and woman, and 𝑚𝑚𝑓𝑎𝑛𝑑 𝑓𝑚𝑓are their married incomes, a necessary condition for a man and a woman to marry is that the income after married is higher than that if they remain single:

𝑚𝑚𝑓 ≥ 𝑍𝑚0

𝑓𝑚𝑓 ≥ 𝑍0𝑓(8)

If 𝑚𝑚𝑓+𝑓𝑚𝑓, the total income achieved by the marriage, is identified with the output of the marriage, a necessary condition for marriage is as below:

𝑚𝑚𝑓 + 𝑓𝑚𝑓 ≡ 𝑍𝑚𝑓 ≥ 𝑍𝑚0 + 𝑍0𝑓 (9) The presence of children is considered as the most reasonable reason for marriage between men and women Children are the only subjects distinguishing single households from married households because sexual demand, care for food and drink, washing and cleaning and many other things can be served by money power, except own children The strong emotion between the two individuals, called

“Love” is also a unique contribution to the purpose of marriage

Cost of marriage, income and the timing of marriage

Another important point is that the gain from marriage is also determined by market opportunities or whether a rise in income encourages men and women to get marriage The expenses of getting married increases to the extent that the own time of man and woman enters into search and other marital costs Many couples can minimize the cost of regular communication and of transfering resource between each other by sharing the same household or living together As a result, this analysis anticipates that acceleration in property income, and a higher level of wage rates, probably enhances the incentive to marry This implication proves a fact that poor people may marry earlier than rich persons but is compatible with the empirical evidence Likewise, this analysis shows that a rise in female’s wage relative to male’s wage, holding the time in household sector constant, would reduce the return frommarriage if women’s wage is lower than men’s wage Since single women work more than married women and single men work less than married men, a growth in wage rate of women comparing to men may reduce the incentive to marry.(Santos, 1970; Freiden 1972)

The traditional family model induces a comparative advantage of women over men in a family because women invest mainly on human capital that raises household efficiency while men are expected to be expert in labor market Hence, “new home economics” suggests a gender-specific pattern of labor in our society and mutual dependence between sexes are major incentives to marry According to this mechanism, the decline in specialization of women due to the increase in economic status (directly resulted by educational expansion) has caused important consequences for marriage First, it can be explained that a successful woman in her career will be a less attractive partner because she cannot focus on her main duty of home production Second, women derive less profit from marriage if they have less need on husband’s income; as a result, economic independence enables women to opt out of marriage since they can afford themselves financial freedom Benard (1972) or Raymo and Iwasawa (2006) proved that high-status women may play an important role in reducing marriage rates

Education and the timing of marriage

“Specialization hypothesis” shows the impacts of education and training on marriage through two causal paths “human capital effect” and “institution effect”

“The institution effect” refers the longer the time an individual stays at school, the lower the possibility for marriage because of three main assumptions proposed by Thornton (1995) Students are not matured enough for adult role, students do not have time for other roles except studying, and married people should be independent in financial aspects Thus, spousal role and studying duty of students are inappropriate Under the division of labor, women tend to leave school when they married since they have to spend most of their time for family As a result, future income of women is also lessened because they have sacrificed their investment in human capital for household work

On the one hand, “human capital effect” has an essential impact on not only marriage, but also fertility and marriage stability at the end of training or education level Despite a higher possibility of stable income in the future, investment on education creates higher opportunity costs for women Therefore, it

is expected that educated women decrease the tendency to marry Moreover, women with financial independence will not gain much from marriage, and thus more of them will not marry at all

If both effects operate, we can find out the combination of them “Institution effect” increases the age

at marriage, and the “human capital effect” also raises the timing into marriage As a result, we expect

a negative effect for both of them, since they move in the same direction

Thirdly, we should consider the channels through which a woman may never get marriage, and investigate how the above theory works in this case “Institution effect” alone cannot let women postpone their marriage forever, and then they will marry as the others lower-educated women do This should not have any special impacts on the proportion of never marry since there is not limit for the age

of marriage, while fertility is limited at a specific age However, there is a debate as women are older, they will be hard to find a potential partners since these partners are all already married, at this time,

“institution effect” seems to be effective This argument may be not enough to persuade, as there are a

lot of men who are staying at school, delayed marriage, too Therefore, we predict that institution effect will not appear in the proportion of never marriage On the other hand, a negative human capital effect shows a strong correlation since highly-educated people are least profited from family formation, so most of them will not marry at all

2.1.2 Theory of Marriage Market

The second conceptual framework I would like to mention is the Assortive mating in marriage markets Because of imperfect information, the process of finding best mates among men and women can be very costly When finding a good match, people tend to set up a minimum level of acceptance, not trying to find a perfect one Those whose conditions are lower than this limit of acceptability will generally not be taken into account However, there is a problem of detecting whether the searching for

a mate occurs or not, young people accidentally start to date when they were in their teens, this time is usually much earlier than the time we assume they are looking for marital partners Specifically, searching for another half of one’s life is often going along with other activities – working, school, entertainment activities and etc A person may not look for a spouse but still find one Given this complication in searching for marital partners’ behavior, the most appropriate strategy may not focus

on whether there is an actual search, but to find out what conditions enables or induce successful searches

Assortative mating

Assortative mating in humans occurs based on a broad array of traits, including social economic, characteristic, educational, residential, traditional, religious, and so on In reality, there are many evidences for assortative mating regarding to altruism Many people in love reveals their similarities in terms of their contributions to public improvement and charities, and generosity can be considered as a proxy for mate choice instead of phenotypic convergence (A Tognetti, 2014) Another evidence comes from the finding of Greenwood et al (2015), which concludes couples also sort for appropriate mate by educational levels and this trend tends to go upward overtime Moreover, assortative mating by

genomic similarities shows its significance in human marriages in the United States In fact, spouses are more genetically identical than two randomly chosen individuals

Equilibrium conditions for Assortative Mating with Monogamy

Identical men receive the same income in an efficient marriage market regardless of whom they marry

or whether they choose to remain single Since marriage with superior women produce larger outputs, superior women receive higher incomes in efficient markets The difference between the incomes of the jth woman and the 𝑖𝑡ℎ woman would be:

𝑍𝑗𝑓-𝑍𝑖𝑓= (𝑍𝑚𝑗− 𝑍𝑚) − (𝑍𝑚𝑖 − 𝑍𝑚) = 𝑍𝑚𝑗− 𝑍𝑚𝑖Where 𝑍𝑘𝑓 is the equilibrium income of the 𝑘𝑡ℎwoman, 𝑍𝑚 is the equilibrium income of men, and 𝑍𝑚𝑘

is the marital output of the 𝑘𝑡ℎ woman and any man Superior women receive a premium that is determined by their additional productivity as wives

If each person is a utility maximizer and chooses the mate who maximizes his utility, the optimal sorting must have the attribute that people not married to each other could not marry without making at least one of them worse off Utility is monotonically related to commodity income; therefore a noncore marriage cannot produce more than the sum of the incomes that its two mates would receive in the core If it could produce more and if any division of output were feasible, a division could be found that would make each better off, thereby contradicting the optimality of the core

The mating of likes or unlike is optimal as attributes are complements or substitutes, because superior people strengthen and support each other when traits are complements and compensate each other when traits are substitutes This theorem also implies that the benefit a woman can obtain from marriage of a given quality is greater for an exceptional man when traits are complements, and is better for an inferior man when traits are substitutes

Positive assortative mating frequently occurs in an efficient marriage market, where superior men are compatible with high-quality women, and inferior men are matched with low-quality women

However, negative assortative mating is sometimes important Maximizing the aggregate output of household commodities is also a feature of an efficient marriage market, where no one can raise the value of his marriage without making others worse off The return from marriage also correlates with appearance, education, brilliance and other traits that have impacts on non-market productivity as well

as market opportunities Holding the market productivity unchanged, the analysis of mate sorting refers acceleration in the value of traits that affect to nonmarket productivity, would cause a higher demand in marriage That’s the reason why less brilliant or less appealing people are going to have a longer time

to marriage than those who are more enchanting and intelligent

The analysis of positive mate sorting proves the fact that the time of seeking for a suitable match will

be extended as educational level increase Women with high-educated level tend to find a man that has similar or even higher educational level than them The case will be more serious if a particular woman has all those traits, like good-looking, intelligent, or well-educated

From the above theory, I expect a negative relationship between women’s expansion level of education and rate of entry into first marriage

2.1.3 Theory of Fertility

Similarly to the mechanism of marriage, women’s fertility is also determined by social factors According to many researchers, the most suitable time for women to have their first child is when they can achieve best result from their plan, including human capital investment and labor market career While man with a bright future plan for career and financial support to his family seems not having any impacts on his wife’s decision of having a child Postpone of demand for children can be explained in the following economic framework (i) women have to scarify their time for taking care of their child instead of being in the labor market Moreover, investing in human capital which results in higher wage

is limited since they have no time for further education These considered as opportunity cost, and women will consider opportunity costs between becoming a mother and investing in their educational level or establish her on labor market Becker (1993) found cost of mother’s time is the most influential

factor to the total cost of children (ii) As soon as a particular woman has her baby, she must leave her current job for a period of time to look after her son or daughter, this means the accumulated experience will remain or even decrease and her skill for this job will also be affected negatively Then, most of them often choose to delay their major responsibility of being a mother

Next, Mills et al (2011) drew a conclusion on negative effect between education and fertility and this idea can be supported by three main literature mechanisms Firstly, the longer women stay in school, the longer the timing of first marriage and first child, this is considered as the responsibility of delay in first child (Blossfeld and Huinink, 1991) Secondly, women who invest more in human capital will have higher probability of achieving success in their career, as a result, the opportunity cost of marriage and giving birth also increases leading to lower fertility (Becker 1991; Blossfeld and Huinink, 1991) Thirdly, more educated women practice the individualism on their career and life more than less educated one, so this discourages family formation and childbearing (Lesthaeghe and Meekers 1987; Liefbroer 2005; Mills et al., 2011)

As we have discussed above, Children are usually not purchased but are the products of marriage between two people from opposite sex; though this process requires a huge sacrify in time of the mothers Each family has their own cost of consumption or different income, the cost of producing and bearing children cannot be the same

Assume 𝑃𝑛 denotes this cost and the cost of Z by πz, the budget constraint of a family equals:

𝑃𝑛n + 𝜋𝑧𝑍 = 𝐼 Where I is full income, given 𝑃𝑛, 𝜋𝑧 and I, the the budget constraint and the marginal utility condition are employed to determine optimal quantities of n and Z:

𝜋 𝑧

The relative price of children and full income are two main determinants of demand for children, thus,

an increase in the relative price of children, in 𝑃𝑛 relative to 𝜋𝑧 , will lower the demand for children and boost the demand for other commodities (where real income is held constant) The relative price of children is affected by many variables, some unique to children, and several of the more important are now considered The evidence over hundreds of years indicates that farm families have been larger than urban families Part of the explanation is that food and housing, important inputs in the rearing of children, have been cheaper on farms

If children can help in household chores, family business or marketplace, then the net cost of children will be reduced Thus, along with the potential gain from children, the incentive for having children also increases In fact, farm families tend to have more children because children are considered more productive for farming than in the cities The contribution of farm children has declined as agriculture has become more mechanized and complex in the course of economic development

Both of these elements have motivated farm families to extend their children's schooling Because rural schools are often too small to be efficient; and it may take too much time and money for farm children

to attend school The cost advantage of rearing children on farms has narrowed, as farm children have spent more time at school Consequently, nowadays, the fertility differentials between urban and rural area have been narrowed in developed countries; while rural fertility is sometimes less than urban area

in some countries

Furthermore, the better opportunities of women in labor market or the increase in the value of time of married women has a significant impact on the relative cost of children The higher income a woman can obtain, the higher opportunities cost of rearing and producing children because cost of mother’s time is the dominant cost of children Indeed, over the last few decades, the steadily increase in earning power of women has become the main explanation for both the large proportion of married women in labor force participation and the major decrease in fertility Since fathers have spent relatively little time on children, the growth in their earning power has no significant impacts on the cost of children and in fact would have reduced the relative cost if children used relatively less time of fathers than

other commodities used Household surveys provide explicit evidence on the relation between the incentive for children and the husbands and wives’ value of time There is a distinguished difference between the effect of father’s value of time and mother’s value of time to the demand for children Specifically, while husband’s increase in wage rate positively related to the number of children, the analysis of mother’s value of time is totally reverse, the higher wage rate a woman earns, the less number of children a woman derives Part of the causation, however, is from children to wage rates, because women invest more in nonmarket sector and less in market skills and men under the burden of financial support have to do the reverse when families have more children However, there does appear

to be significant causation from the value of time of wives to their demand for children

2.2 REVIEW OF EMPIRICAL STUDIES

There are various studies concerning the impacts of education on timing to marriage and first birth For example, a comparative study was conducted by the data from U.S and Germany national surveys (Josef Brüderl, Andreas Diekmann 1997) investigated whether higher level of education only delays marriage or if it also reduces marriage intensity Also, educational effects for the case of United States and West Germany will be compared Log-logistics distribution with transition rate to model age-dependent marriage probabilities is applied to estimate the marriage rate And to separate human capital effect and institution effect, authors also used the model of generalized log-logistics The dependent variable here is the waiting time “t” until a woman married, which is measured by marriage age minus the minimum marriage age, from this measurement, author can infer the rate of first family formation – H (t) To see how this rate fluctuates, the model includes some social variables such as years of education, professional prestige of father, religion, place of residence and number of siblings etc Based on the aim of the study, education variable is clearly defined and divided into some different

variables, specifically: Years of education (EDUC) is separated by level, such as primary, secondary,

high-school, college and university and EDUC is measured by the numbers of years which a particular

woman can achieve that degree Lower education is a dummy variable, and code 1 refers people with

level of education is less than nine year or illiteracy Years of education after a marriage process is the

time of training after minimum marriage age, and is formulated by EDUC + 6 which is the minimum marriage age, if a woman stops studying before minimum marriage age then this index is equal 0 The purpose of this variable is to separate the institution and human capital effect Professional prestige of father is the professional of women’s father when a woman was 16 years old Religion is defined as a dummy variable of Catholics, Jewish and others religion The reference group is Protestants and non-religious people Place of residence refers to whether women live in rural area or city when they were

16 Number of siblings is the number of brothers or sisters the respondents have

As a result, in both countries, a negative educational effect was found for both genders, or the longer a woman or man spend at school, the longer until their first marriage Besides, women present a strongly negative human capital effect over all birth cohorts and it is consistent with family economic theory For U.S, the effect of lower education is significant for longer time of marriage and the median for them is 2 years later, but it shows no effects for the case of Germany The meaning of years of education after a marriage process is amazing since it reveals one more year of education delays the maximum rate of marriage of 5 to 11% There were vulnerable education effects in East Germany, since this is a socialist nation with solid family policy interventions The growth in number of people involving in tertiary and secondary school has postponed marriage as well as leading to higher proportions of never marrying Author also concluded that increasing level of education cannot be the sole explanation for the observed marriage decline because more and more people decided not to marry even controlling education

Another study of first marriage determinants was conducted by Wong (2005), who applied Hazard analysis to measure the rate at which marriage takes place This dependent variable varies with “t”, which is the age at first marriage or current age if respondents have not married yet Wong also used 2 types of independent variables of family background covariates and educational or career to explain this relationship For education variable, Wong divides it into 3 categories: primary and below,

secondary (base-line) and post-secondary Job position is also proxied by three categories based on Hong Kong statistics department (1997) managerial and professional, clerical, primary and unskilled

By adopting the integrated approach, this paper reveals the more investment in human capital, the longer time to marriage for women More specific, women finish post-secondary; have a lower risk of enter into marriage than those who only have secondary-degree This effect is similar for the result of job position, women in a higher social class account for a small proportion in getting marriage Also, women with traditional attitude tend to marry earlier than those who don’t under the same society by the analysis of birth place

Consistent with the results of Wong (2005), a recent paper of Abalos (2014) focuses on trends in marriage of both men and women in Philippine, has used the data from National Demographic and Health Survey in 2003 and 2008 for analysis According to Abalos, marriage is considered as a formal marrying convention between 2 people or living with each other and cohabitation Age at first marriage

or timing of living with first spouse is dependent variable and this study practices the event-history analysis or survival model in measuring the time until a woman marries At time of survey, there are many women still remaining single, so the data set may consist of censored cases That’s also the reason why author has to use the model of hazard rate function to take into account the effects of covariates – education, ethnicity, religion, place of residence, birth cohort and etc In this model, education variable is measured by the highest level of education that a woman obtained, and separated

by primary level or lower, secondary and higher than secondary school Birth cohort is women’s years

of birth, place of residence is where a woman was born, classified by rural and urban area As we have mentioned above, Abalos’s findings are quite similar to the consequences of Wong, educational expansion and place of residence show a strikingly impact on marriage postpone, especially for women One more interesting point of this paper is the increasing in the number of living together without marriage trend among men and women in Philippine, it must be an inaccurate result if the data ignores this case

Along with the delay of first marriage, postpone of women’s fertility cannot be prevented, especially Asian countries with traditional culture life Ida & Albert (2015) investigated the determinants of age at first marriage by applying Log-logistic model to estimate the probability of not having first birth until time t In this model, residence at first birth (whether a woman was born in rural or urban area), highest level of education (the most valuable degree a woman can achieve), religion (a group of popular religion that people practices such as Buddhist, Catholicism, Islam), wealth of index, ever have an abortion, type of contraceptive used, age at first marriage and first intercourse are taken into account as explanatory variables Among nine covariates, only age of a woman at first marriage, educational level and type of contraceptive used are significant, wealth level of a woman is also significant at a lower degree, while the others present no impacts on women’s first fertility

Glick et al (2015) used continuous-time hazard model to estimate age at first birth with an arrays of independent variables, including women’s parents educational level, mortality, asset index, religion and ethnicity of head of households, women’s education, birth-cohort The definition of women’s education

is the same the previous papers, it is the number of year a woman can achieve her highest educational degree One more unique point of this study is the analysis for the effects of time since marriage on the timing of first birth; as a result, Glick found that the risk of having children increases from 0-3 years after marriage and then decreases This study also reveals the importance of schooling on women’s fertility, specifically, one added year of schooling accounts for 0.5 year of delaying fertility

Since there are a number of researchers investigating in this area, this paper tends to contribute to shed

a light for the case of Asian countries – Vietnam

I collected the data of 505 respondents, including 32 male and 404 female, but there are only 304 female respondents answer all the questions Therefore, my final database is a random sample of 304 women between 18 and 66 years old coming from both provinces and cities in Vietnam Specifically, there are 304 women gave full answers for information on first marriage and 303 respondents fully answer for the case of first birth

3.2 METHODOLOGY

Following the method applied by most of previous papers (see for example, Blossfeld el al 1989; Blossfeld and Huinink, 1991; Tuma & Hannan, 1984), I use the method of survival analysis to model the behavior of entry into first marriage and motherhood Let the event under consideration be the first marriage or motherhood, the survival function, which indicate the rate at which the woman remains single (for the case of marriage analysis) or not having any children (for the case of motherhood) at time 𝑡, is

𝑆(𝑡) = 1 − 𝐹(𝑡) where 𝐹(𝑡) is the cumulative distribution function for the event (marriage or childbirth) to occur The distributions usually assumed are exponential, Weibull, log-normal and gamma The hazard function ℎ(𝑡) which shows the instantaneous rate at which the event (marriage or childbirth) occurs is

ℎ(𝑡) = 𝑓(𝑡)

𝑆(𝑡)Where 𝑓(𝑡) 𝑖𝑠 the probability distribution function is associated with 𝐹(𝑡) and 𝑆(𝑡) is the survivor function defined above

In addition, I also analyzed some crucial factors affecting to timing of first birth for women, in differenct levels, including individual aspect, family aspect, social, economic, internal and external aspect as the below graph

Age at first Child birth

Early Adolescent factors, e.g

Peer pressure, lived with both

parents, age at first marriage

Cultural and Social factors e.g norms, practices, taboos, international factors

Age at first sex and marriage

Two edged factors e.g Education, career plans

Background factors

e.g Religion, Residence,

Race, Parents Education,

Economic status,

Employment status

To analyze the impacts of education and career on the entry into first marriage and childbirth, we model the hazard function ℎ(𝑡) as a function of education and career related variables, together with other control variables, with the assumption of the exponential distribution

ℎ(𝑡) = 𝑒𝛽𝑋And Cox Hazard Model with the equation as below

ℎ(𝑡) = ℎ0(𝑡)𝑒𝛽𝑥where ℎ0(𝑡) is baseline hazard function

3.3 VARIABLES’S DEFINITION

The variables used in this analysis are defined as follow

+ Dependent variable is the age at first marriage or first entry into motherhood (Measured by years),

and event-history analysis is applied to find out the maximum time until a woman marries or have her first baby related to the below regressors (or covariates)

+ Other independent variables (time-invariant covariates) include

1/ Father’s social class or Mother’s social class with the set of class categories: unskilled and skilled

manual workers (1); unskilled and skilled administrative and service workers or Clerical (2); professionals and self-employed (3) Unskilled and skilled manual worker’s class is considered as reference category (Table 3.2)

2/ Father’s education and Mother’s education is divided into four categories, in which Secondary is

base line group, vocational training (1), college (2), university or higher (3) (Table 3.2)

This is because our country suffered from such damage from two devastating wars in the past, the opportunities for women’s father to access to advanced education is limited Second, Vietnam is a developing country with the traditional economy of agriculture Although I collected data for both Father’s social class and Mother’s social class, I prefer to use the variable for father since Vietnam has

a long tradition of Feudalism; male children had priority access to the available educational resources

than female Additionally, we are analyzing a generation that practiced the traditional division of labor which fathers are responsible for earning for living outside the home and mothers’ duties are working within the household, thus father’s education would perhaps be a better appraisal of parents resources

in this case However, I still applied both Father’s and Mother’s social class into the model to see whether it has any distinguishes

I expect a positive relation between high father’s social class and marriage delay, because parents of these social classes tend to invest more in their children’s future through education Consequently, the more educational opportunities they offer to their children, the longer their children stay at school, leading to an increase in marriage age This is consistent with the “institution effect”

3/ Type of residence (town or country) a dummy variable measures how environmental surroundings

affect to marriage behavior The residence dummy variable is an indicator to observe whether women from rural area have different behavior in family formation than those from urban area, and country is seen as reference category It is obviously to say that women from rural area marry earlier than those in urban area, especially in Eastern countries This is also a popular result from the findings of Huinink (1987) or Huinink and Wagner (1989) Moreover, this variable is also found to be significant for women’s early birth The better access to advanced infrastructures and facilities in urban area could induce the disparity between rural and urban life For example, urban area creates more jobs offering for women to afford their further study which may a contribution to postponement of first birth Additionally, parents in urban area tend to have better guidance and motivation for their children to commit in school attainment, thus, in this situation; women from rural are probably forced to have children earlier (Table 3.2)

4/ The number of siblings is collected by data of each household and considered as a factor which

influences the time of entry into first marriage, since the larger the family, the higher probability of low level of education, so they incline to get marriage earlier as well as having children (Table 3.2) Firstly, women from larger families are disadvantaged from educational investment, since very few couples of

parents can fully focus on their children future’s career Secondly, these women are ready for housewife and motherhood as they are required to take care of their siblings when parents are at work This prediction is based on the findings of earlier studies, Huinink (1987), Huinink and Tuma (1988), Huinink (1988b)

5/ Birth-cohort Different birth cohort is applied to predict the differences in marital timing among

generations (Table 3.1) The tremendous socioeconomic change that occurred since 1975s has probably alternated the marriage attitudes of women as well Overall, we expect a positive relation between marital time and earlier birth cohorts, in other words, women born in earlier birth cohorts may marry at younger ages than those in later birth cohorts This variable is another set of dummy, each representing 4-year birth group of 1992- 1998; 1987-1991; 1981-1986 and 1950-1980 (reference group: 1992-1998)

Our objective was to see if there is an independent cohort effect after combining the other determinants such as job position and education Finally, we realized that women from younger birth cohorts have entered marriage at a continuously later age The remained question was whether this cohort effect, which represented as an indicator of changing attitudes toward marital timing, was still significant even after we had included other explanatory variables in our model

6/ Chief income earner It is obviously understandable that a woman with burdens of financial

obligation to their parents or even siblings may marry at later age (Table 3.2) As a chief income earner

in her natal family, women’s behavior is also different since they have higher responsible soul and they will not marry until there are alternative arrangements to support her family This is a dummy variable with based line is group of women who are not chief income earner in her family

7/ Ancestral worship is an indicator of respect for the ancestors (deceased parents are also included), as

well as a considerable mechanism of transferring traditional values or cultures through generations (Table 3.2) Given the fact that women who regularly worship ancestors are considered as holding strong traditional values, and they can easily accept early marriage and childbearing For this reason, it

is reasonable to investigate whether there is any difference in the timing of entry into first marriage between women who worship ancestors and those who don’t

8/ Woman respondent’s birth order in the family Our country’s culture is mostly affected by the

Confucianism, in which children in the family often have to wait until their turn to enter into marriage life in accordance with birth order Holding the others factors constant, women who have many brothers or sisters are obviously have to spend more time in “queuing”, the case will be worse if they are from a large family Birth order in our data set is divided into three groups, one group of eldest child, another is youngest child and third group is others, reference group is women who are eldest children in her family

9/ Job position refers women’s prospects and achievement (Table 3.1) This variable is categorized by:

unemployed and unskilled or skilled labor, clerical, managerial and professional Unskilled labor is a group of workers, blue-collar, farmers as well as service jobs; which is used as reference group Clerical was mainly classified by office workers Managerial and Professional includes administrative position, managers and those who run their own shops or companies and do a small business as shop owners Based on our assumption, it must take longer time for those have greater skill requirements and well-purchased career to get marriage Particularly, with this classification, I predict the group of Managerial and Professional women will be have the longest time to marriage, and unemployed and unskilled labor group to marry the earliest

10/ Ethnic We expect some unique ethnic will have different marriage and childbearing behavior

comparing to the majority (Table 3.1) In this paper, I divided into two separated groups of ethnicity including Kinh, Hoa and others which accounted for highest proportion in our country’s population I expect to see there are something differences in timing to marriage and have first birth between these groups

11/ Income at first marriage or current income if has not married yet This income is asked directly

from the respondents and the income level is separated into multiple groups, including lower than 5

million VND per month, 5 to 7 million VND per month, 7 to 10 million VND per month, 10 to 15 million VND per month, 20 to 25 million VND per month, 25 to 30 million VND per month, 30 to 35 million VND per month, 35 to 40 million VND per month, 40 to 45 million VND per month, 45 to 50 million VND per month and over than 50 million VND per month (Table 3.2) Then we used the average income for month as value for each group, for example lower than 5 million per month group will receive the value of 2.5 million VND per month However, in regression model, this variable is considered as continous variable, and the median value of each income range is taken into account For example, woman who has monthly income from 0 to 5 million VND per month, will give the value of 2.5 million VND per month

About the meaning, we hope to see the result of the higher the income, the longer the women marry or have their first child

12/ Women’s education This variable is classified by three categories, including college or lower,

university, master and higher And this is also measured by years of schooling a woman needs to achieve a particular degree Specifically, Primary = “5 years”, Secondary = “9 years”, High-school =

“12 years”, Vocational =”14 years”, College =”15 years”, University =”16 years”, Master and above is more than 16 years The number of schooling years is asked directly in the questionnaire In these categories, lowest educational group will be considered as base line for comparison (Table 3.1) We expect the higher the educational level, the longer time to marriage or motherhood

13/ Women’s promotion in her career This is a set of dummy with 1 is coded as ever get a promotion

and 0 is understand as never get a promotion, which is used to measure the career opportunities of women We expect a negative relationship between promotion and time at first marriage or first birth (Table 3.1)

14/ Women’s Job movement a dummy variable with 1 is coded as ever change job before and 0 is never

change job This shows the stability in income or women’s career (Table 3.1)

15/ Living arrangement is set as dummy variable with 0 for not living with parents and 1 for vice versa,

we expect living arrangement before marriage may affect their decision on marriage Particularly, women live with their own parents usually delay marriage or even first birth since they may have to take care of their parents or siblings (Table 3.4)

3.3.2 First-birth variables

Beside all the variables in the equation of first marriage are reused, there are some new variables in the equation of entry into first birth, including the below variables:

1/ Ever terminated a pregnancy This is a type Yes/No question and is coded 1 for Yes and 0 for No in

the model (Table 3.4) The negative relation between early births and having terminated a pregnancy is predicted, since women with the experience of abortion may understand clearly the risks of this operation, and they probably earlier accept to give birth right after marriage

2/ Wealth index this variable is self-explained and is classified by property (Table 3.4) Respondents

are asked a question of whether they have a valuable car at time of marriage to evaluate the wealth status And I only separated this variable into lower income and higher income groups According to previous empirical study (Chen and Morgan), the effect of poverty on time at first birth is needed to be considered Women from poor households with low level of education tend to marry earlier than those from more educated and wealthier families

3/ Age at first intercourse is encountered, since childrearing cannot occur unless one engages in sexual

activities which eventually cause pregnancy (Table 3.4) The date of first marriage is the conventional marker of the beginning of hazard of pregnancy However, sexual activity is not constrained to marriage and many women still have children before the date of marriage Thus, age at first intercourse and date of first birth are suitable proxies of the beginning of sexual exposure than the date of first marriage

The positive relation between time to first birth and age at first intercourse represents that later beginning of sexual intercourse extend the waiting time to first birth

4) Knowledge about contraception separated by two categories, unknown is 0 and known is 1 (Table

3.4) If a woman has been informed how to prevent from pregnancy, it is reasonable to expect that she can control their time to motherhood

5/ Contraception type This variable is divided into: Never used, Condom, Contraceptive pill, set a coil

and others (Table 3.4) Women with modern types of contraceptive is expected to delay their time at

first birth, because the more modern technique, the lower risk of unexpected pregnancy

6/ Age at first marriage The respondents are asked at which age they married, if they have not married

yet, their current age will be taken into account for this variable (Table 3.4) Women married at later age are expected to have children later than those who married earlier Although, we construct this variable, it is not feasible to put it into our equation since there are many censored data at time of survey for those who remained single until time of survey and their information is considered incomplete

7/ Educational level of partners Educational level of partner is expected to be an advantage for women

to have children earlier, so it needs to be taken into account (Table 3.4) This variable is classified from primary to professor level, which is measured by the number of schooling years Primary level is used

as reference group From a theoretical point of view, men with good job, high level of income and career need more time to prepare for their marriage life, Oppenheimer (1988) Similar to Age at first marriage, we cannot observe the partner’s educational level for those women who have not married yet, that’s the reason why this variable does not appear in our regression model

8/ Housework This variable aims to investigate whether a woman hold traditional values or not, since

those who usually do housework are supposed to be ready for wives’ or mothers’ role than those who don’t Therefore, this variable is coded as 0 for those who do not often do their housework and 1 is for another Obviously, 0 is considered as reference group (Table 3.4)

Table 3.1 First-Marriage Variables Description (Education and Career Variables)

1 Women’s education

(0-Base line)

0- College or lower 1- University 2- Master and higher

Reference group is expected to marry earliest compared to other group

2- Managerial and professional

Reference group is expected to marry earliest

4 Birth-cohort

(0-Base line)

0- 1992-1998 1- 1987-1991 2- 1981-1986 3- 1950-1980

Reference group is expect to marry earliest

5 Working experience

(years)

Measured by years at work The longer the time at work, the

longer waiting time until first marriage

Table 3.2 First-Marriage Variables Description (Social Background Variables)

1 Residence

(0-Base line)

0- Rural area 1- Urban area

Women from rural area is going to marry earlier

2 Father’s social class

(0-Base line)

0- Unskilled and skilled workers 1- Clerical

2- Managerial and Professional

Women with lower father’s social class tend to get marriage earlier

3 Father’s education 0- Secondary

1- Vocational training 2- College and University 3- Master or higher

Women’s with well-educated father tends to have higher level of

education, and then later marriage is expected

4 Mother’s social class 0- Unskilled and skilled manual

workers 1- Clerical 2- Managerial and Professional

Women with lower mother’s social class tend to get marriage earlier

5 Mother’s education 0- Secondary

1- Vocational training 2- College and University 3- Master or higher

Women’s with well-educated mother tends to have higher level of

education, and then later marriage is expected

7 Chief income earner in

her family (0-Yes)

9 Ethnic (0 – Base line) 0- Kinh

1- Chinese and others

Expected to see difference in time to marriage between ethnicity

11 Women’s birth order

(0-Base line group)

0- Eldest child 1- Youngest child 2- Others

The lower the birth order, the more waiting time until first marriage

1- Buddhist 2- Christian

Expected to see difference in time to marriage among Religion

Table 3.3 First-Birth Variables Description (Education and Career Variables)

1 Women’s education at

first birth

(0-Base line)

0- College or lower 1- University 2- Master and higher

Reference group is expected to have children earliest compared to other group

2- Managerial and professional

Reference group is expected to give birth earliest

4 Birth-cohort

(0-Base line)

0- 1992-1998 1- 1987-1991 2- 1981-1986 3- 1950-1980

Reference group is expect to have children earliest

5 Working experience at

first birth (years)

Measured by years at work The longer the time at work, the

longer waiting time until first birth

7 Income at first birth

Table 3.4 First-Birth Variables Description (Social Background Variables)

1 Ever terminated a

pregnancy

0- No 1- Yes

A woman with answer “yes” is going

to give birth earlier

2 Wealth index (by

asking a direct

question of property)

0- No 1- Yes

A woman from poorer family tends to have her first birth earlier

3 Age at first

intercourse

0- Not yet 1- Lower than 18 years old 2- 18 to 22 years old 3- Over than 22 years old

The earlier the time at first intercourse, the sooner a woman has probability of first birth

4 Knowledge about

contraceptive

0- Yes 1- No

Women with knowledge of contraceptive can prevent her from unexpected pregnancy better

Therefore, they are going to have children later

5 Contraceptive type 0- No use

1- Condom 2- Contraceptive pill 3- Set a coil and others

We would like to see the differences among these kinds of pregnancy precaution

6 Age at first marriage Asked directly the age at which

respondents entered into marriage If not married yet, then their current age will be applied

The sooner time of marriage, the sooner time of first birth

7 Partners’ Education Number of schooling years The higher level of education, the

shorter time of first birth

8 Housework at first

birth (0-Base line)

0- Never do housework 1- Do housework monthly 2- Do housework daily

The more industrious a woman was, the earlier time to first birth is expected

Table 3.5 First-Birth Variables Description (Social Background Variables)

1 Residence at first birth

(0-Base line)

0- Rural area 1- Urban area

Women from rural area is going to have children earlier

2 Father’s social class

(0-Base line)

0- Unskilled and skilled workers 1- Clerical

2- Managerial and Professional

Women with lower father’s social class tend to have children earlier

3 Father’s education

(0-Base line)

0- Secondary 1- Vocational training 2- College and University 3- Master or higher

Women’s with well-educated father tends to have higher level of

education, and then later first birth is expected

4 Mother’s social class

(0-Base line)

0- Unskilled and skilled workers 1- Clerical

2- Managerial and Professional

Women with lower mother’s social class tend to have children earlier

5 Mother’s education

(0-Base line)

0- Secondary 1- Vocational training 2- College and University 3- Master or higher

Women’s with well-educated mother tends to have higher level of

education, and then time to first birth

7 Chief income earner in

her family (0-Yes)

9 Ethnic (0 – Base line) 0- Kinh

1- Chinese and others

Expected to see difference in time to fertility between ethnicity

11 Women’s birth order

(0-Base line group)

0- Eldest child 1- Youngest child 2- Others

The lower the birth order, the more waiting time until first birth

13 Religion

(0-Base line)

0- Non-religious 1- Buddhist 2- Christian

Expected to see difference in time to fertility among Religion

Chapter 4

Empirical Result

4.1 EMPIRICAL RESULT

4.1.1 Statistics

Table 4.1 Summary statistics of First marriage

Table 4.1 shows detail on income, work experience, number of siblings and years of schooling of women in the data set Overall, the average income of 304 women in our sample is about 8.748 million VND per month with minimum and maximum amount of 2.5 million and 50 million VND per month respectively Average working experience of these women is over 6 years, with a number of females still does not enter into labor market Number of siblings ranges from 0 to 10, since this data is collected from women across cohort, and it is not surprising that previous generation had more children than modern women Years of schooling is the number of year that a woman really needs to finish a particular level of education Maximum 24 years of education in this table refers to the fact that this woman has spent 24 years on investing in her education