Age effects on mothers’ education with respect to the first born child

(1)

1

Thesis

Name: Coen van den Brink

Student number: 10775471

Specialization: Economics

Field: Labour Economics

Number of credits: 12 EC

Title: Age effects on mothers’ education with respect to the first

born child

(2)

2

Statement of originality

This document is written by Student Coen van den Brink who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision and completion of the work, not for the contents.

(3)

3 Table of Content

Table of Content 3

Abstract 4

I. Introduction 5

II. Literature review 7

- Literature in the past 7

III. Methodology 11

- Hypotheses 11

- Data used 12

- Estimations 13

- Modifications of the data 16

IV. Results 16

V. Discussion 20

VI. Conclusion 21

References 23

(4)

4

Abstract

In this paper, the effect of the age of mothers when they get their first child on her education with a link to her labour market outcomes is researched. So far, a lot of research has been done on the effects of having children on their parents. This thesis reviews the existing literature and subsequently examines the relation between the age when women get their first child, and her education and future career. Data of women born between 1950 and 2000 is used in the OLS regression estimates. Significant findings are that there is indeed a small positive interaction between these variables. When a woman chooses to have her first child later in her life, this will have a positive effect on her education and with that her labour market outcomes in the future. It is also found that this effect has not changed during the past half century, however this outcome is not significant.

(5)

5

I. Introduction

There are several causes of the financial backlog position as well as the position of women on the labour market. Women with the same qualifications as men are mostly underpaid, also called the gender wage gap. Although this gap has been shrinking during the past decades, women’s career steps are nowadays still smaller than men’s. The major cause is the 24/7 work culture, which punish people who have responsibilities outside the working place such as parenthood. Often, this parenthood boils down on the female in the family. Some even promote to change our social system back to where only one parent ensures an income for the family, because of the lack of time of the parents and thereby the low quality childcare (Alderman, 2001). In the past decades, much more research has been done not only on the effects that having and nurturing children have on the careers of their parents, but also on the educational levels they achieved. However, previous studies dealt with characteristics related to the children, such as the number of children and whether they were hard to raise. Also, much research on the reverse effect of schooling of the parents on the schooling of the children has been done, which in most outcomes show a positive effect (Angrist & Evans, 1998). However, in these studies the characteristics and the effects on the positions of the mothers were not taken into account. This study even though, deals with specific age effects of the mother and thus the actual time she will have to provide her own education. With that a comparison of the situation of several participants is done and a conclusion has been drawn with the found results. Motivation behind this, is that this research distinguishes itself from previously done research, because it focusses on the age of the mother when getting her first child and not on number of children or schooling.

Research in the past focused on emotional but also tangible effects which have been written about. Most outcomes explained negative health and social issues for both the mother and the child if the child grew up in a young motherhood. A frequently common consequence of early motherhood for children is single parenthood, which is disadvantageous for growing up and bringing up (Angrist & Lavy, 1996). Furthermore, many research programs studied the effect of children on careers of the parents. Some of the research previously done focused on theoretical information of labour market participation of women in the past years (Cools et al., 2017). On the other side, a lot of empirical research has been done on subjects where having children may influence

(6)

6 the future of the parents. For example IV, Estimations of “childbearing on labour supply” (Angrist et al., 1998) and “Effects of childbearing at an early age on that children” (Angrist and Lavy, 1996). Many researchers conclude that the dilemma is that in the last decades, education and career achievements are increasingly improving and becoming more important to women. Nonetheless, no solutions are given for the fact that these women most times took care for their children, but now also work on their careers, while these children still need the same care nowadays compared to the past. So, in a lot of cases the child care is a problem, which is a pretty expensive burden that people want to bear for their children. Another problem is that this daycare is at such a low level, that people try to find other solutions for the care of their children (Beets, 2004).

On the other side, however, getting children at a later age poses health risks for the mother as well as for the child. Situations of frequent occurrence are that parents do not have enough time for their children and bring them to mostly low quality childcare organizations where children do not get the care they actually need. Having said this, the perspective that parents have in their education and, with that, their career does play a big role in raising children (Morris, 1989).

In this study, the following two questions will be answered: whether there is a significant effect of the mother’s age when she gets her first child on her educational level, her labour market outcome and whether this effect has become smaller or larger in the past decades. Data of the Organization for Economic Co-operation and Development (OECD) measures the educational level in years of schooling, and the data set provides information about the age of the mother when she gets her first child. This dataset contains information about 1,352 females in the Netherlands about their age and their educational level in years of education as well as the age of their first born child. Based on these data, we can compute the age at which the mother got her first child. A regression will be done on the age of the mother when she gets her first child, the independent variable, and the education of the mother, the dependent variable. It is assumed that the sex of the children does not have an effect on the education of the mother. For answering the second question, the data will be split up into age classes and the effect will be measured. After this, the different classes will be compared, and it will be determined whether the effect has grown or shrunk in the past decades (OECD, 2011).

(7)

7 The data set contains merely Dutch women, so it is also assumed that all women included in the data set got an almost equal chance in achieving the educational levels they aim at. In the Netherlands, there is almost no cultural pressure in getting children at an early age, so that factor can be excluded. Because generally getting children has certain impact on someone’s life, it will also affect achieving certain educational levels if a woman becomes pregnant in an early stage of her life. As well as on education, children also have an effect on the career of their parents, especially on the mother’s because most times nurturing of the children is taken care of by the mother. She will thus have to interrupt her learning process or career.

The findings in this thesis show that there is a significant effect of children on education of the mother. The results suggest that getting children at an early age affect education in such a serious way that the mother is not able to go to school or university anymore. She has to stop her education and at the same time, despite her educational level being lower than the level she could achieve based on her intellectual capabilities, ensure that it is financially possible to take care of her children. So, there will probably be a positive relationship between the age of the mother when getting her first child and her education. Getting children at a later age means having more time to spend on education and after that, on her career. Knowing this, we can link educational achievements to earnings, where also a lot of research has been done into.

The following will be discussed in this thesis. First a literature review with all the research done previously and empirical studies in the past is presented; subsequently, the further sections will discuss the characteristics of the data, the results found, the limitations of the estimations will be discussed (chapter V) and conclusions will be drawn (chapter VI).

II. Literature review

- Literature in the past

Beets (2004) researched, as a result of the anxiety of the Dutch Secretary of State on the coordination of emancipation, the timing of the first child with respect to the future perspective of the mother. He discusses considerations of parents whether starting a family and thus the effects of early and late fertility. Shortly after the second World War the average age of getting children for women was around 24 years and

(8)

8 20 years later this has increased to 29. According to Beets there are several explanations for this risen age. The first and biggest cause he mentions is the improvement in education. This gives women the opportunity to build up educational levels which help them further in their career for the rest of their life, but, with this, in their early ages they do not have time for children anymore. Another reason frequently given for not having children at early ages is that (mostly) high educated woman have not found their appropriate partner yet. A third argument, according to Beets (2004), is the emancipation of women or feminism, which makes women more independent and therefore gives them more time for other important things in their life than children. The conclusion is that most women consider the age between 25 and 29 years the best age for getting children; getting children after this age would increase the probability of health problems and Beets (2004) gives several advices for policy adjustments such as reducing the numerous possibilities for parents to participate in the labour market. Moreover, this does not only have to apply to women but to all parents. But the most important correction he instructs is the improvement and cost reduction of day-care.

Holmlund, Lindahl and Plug (2011) examined the effect of parent’s schooling on the schooling of the children. Based on the Swedish data set that is used it can be concluded that there certainly is an effect of educational level from generation to generation. This paper can be relevant for my own research, because apparently the educational level of the parents and especially the mother’s educational level and later career, which I am researching, are factors that have influence.

Angrist and Evans (1998) did an IV estimation on the effect of having children on careers of high and low educated parents. The IV estimation is used because of exogenous effect of fertility and sibling sexes and the researchers want to rule these out. Their investigation starts with the educational level of the mother and measures this in graduations in schooling. The results show that there is indeed an effect of having children on the labour market outcomes of the parents, but the most important outcome of the research is that this effect is almost nil for highly educated parents, whereas this effect does appear for poor and low educated mothers. Another result indicates that fathers scarcely adjust their labour market behaviour. The opportunity costs of the mother who does not earn money for the family are absorbed by the family. Conclusion is that in the time the research was done by Angrist and Evans, the number

(9)

9 of women joining the labour market has increased a lot. However, the labour market behaviour of men has not changed.

Cruses and Galiani (2007) find that the result found by Angrist and Evans (1998) can be generalized from developed countries to developing countries. The researchers generalize the results qualitatively and quantitatively to labour supply in both Argentina and Mexico with the so called “mixed sex siblings preference” IV estimator that was also used in the research of Angrist and Evans (1998). Their conclusion is that the estimates of Angrist and Evans (1998) from the US can be generalized quantitatively as well as qualitatively to developing countries. Which means that if fertility of women is higher, their educational levels are lower and there are less facilities for taking care for children.

In addition to Angrist and Evans (1998), Cools et al. (2017) investigate long term effects of family size on labour market outcomes for the careers of Norwegian parents’. Parents prefer a combination of at least one boy and one girl in the family, so if two children with the same sex are born, the effect will be that the parents want to get another child. The researchers think that this effect will bias the results and use for this an IV estimation called sex composition. Mothers mostly have a break when bearing their child and work less when they take care of their children. The results that are found, show different outcomes for the short, medium and long run. In the short and medium run, the number of children give a negative effect that is significant on labour outcomes of the parents. However, in the long run this disruption is brought back. So, it is often thought that parenthood causes persistent career effects for parents. However, significant results on the research field show that this is not the case.

Angrist and Lavy (1996) studied the effect of becoming a mother at a young age on their children’s health and schooling. The common thought is that young mothers cannot take proper care of their children for two reasons: the first problem is money, because young mothers cannot finish their schooling, their income will be low, and mothers cannot take care of their children well, and secondly, time is a problem. Because young mothers often also are single mothers and in some cases also have to study, they do not have enough time to look after their children. The researchers used two data sets and with this measured the effect of the age of the mother on their so called disability and grading repetition of the child. The data also includes information on whether the father of the child is present. Angrist and Lavy (1996)

(10)

10 conclude that this there is a small effect of the presence of the father in the family on the health of the child. However, on the grades of children, it has a much stronger impact. The overall conclusion is that childbearing at a young age is not that substantially bad for children’s health and schooling, but the presence of the father has a big effect on the two variables.

Butcher and Case (1994) documented the effect of sex composition in a family on the education and earnings of the siblings. The outcome was that women with brothers get more education compared to women with just sisters. More research has to be done to give an explanation for this. They also concluded that the family structure is a factor that has influence on education of the children. In short, this paper leaves a lot of open questions.

Ambert (1993) wrote a book on the effect of children on parents in various aspects discussing intangible, such as mental sides, but also tangible effects, such as education and earnings, of children on their parents. In most of the examples she uses, children that are hard to raise are mentioned. In this book all the costs, but also additional benefits are discussed.

Finley (1998) discusses the conclusions made by Morris (1988) about the perceived effect of parental age on the quality of parenting. Participants were students from a large urban University. The findings of Finley (1998) deteriorated the work of Morris (1988) with the facts that he noted that there was no significant effect on the quality of care of the mother and maternal age. For the fathers, however, there was a significant effect. Finley (1998) concludes that it is not possible to provide a clear statement about the quality of parenting and the age of the parents. This adds something to this thesis, because it does not debounce the effect that is expected. It is tried not to let the characteristics of the child suffer from the results that may occur from this research.

Van Noord-Zaadstra et al. (1991) did a cohort study examining the age under which women became less fertile. 751 participated in a donor insemination program and were married to a husband who was infertile. The outcome was that the age at which fertility seriously decreased was at the age of 31. However, this can be atoned by continuously fertilize the woman.

Most of the research that has been done gives advice on whether it is wise to get children at a certain age. Much more literature discusses the effects that characteristics of parents have on their children. On the one side, mothers that are

(11)

11 young, so between 18 and 25 years old, are more fit and the chance of getting healthier children is higher compared to older mothers, which can cause health risks. The disadvantages of young motherhood are the financial problems that frequently occur and one-parenthood that appear sometimes. Advantages of becoming mother at a later age are that in most of the times, the mother has had time to finish her education and thus has a good job where she earns enough money to take care of her children. Another important factor that plays a big role is the presence of the father. If there is a possibility to divide the nurturing of the child, the female in the family will have more time for her own education and career. This makes ‘presence of the father’ a variable that has impact on the dependent variable which is the education of the mother.

III. Methodology

- Hypotheses:

After having reviewed all the research previously done, the Hypotheses of this thesis can be formed. These hypotheses will be examined in the research done in this paper: Hypothesis 1: There is a positive effect of the age of the mother when she gets her first child on her own education.

Hypothesis 2: There is a positive effect of the age of the mother when she gets her first child on her career.

Hypothesis 3: This effect has increased from the 1950’s till now and will probably increase further in the future.

The intuition behind the first two hypotheses is that if the mother has more time to spend on her own education and does not have to spend this time on the care of her child, she will actually achieve a higher education. Even though Dutch data is used and the Netherlands is a fairly liberal country, the mother will not be compensated in her future education for the time she ‘lost’ in taking care of her child. The third expectation can be intuited by the fact that in the past, the mother was participating less in the labour market compared to now. It is thus expected that that effect will become stronger. Whether the hypotheses can be confirmed will be explained in the

(12)

12 results. In the next section, the data found and the method of research will be discussed. A linear regression will be used on the data, so any other relation between the numbers cannot be found in this research paper.

- Data used

The empirical part of this thesis starts here. As already said, a dataset of the Organization for Economic Co-operation and Development (OECD) is used with a lot of data on citizens from the Netherlands with country code 528. The Netherlands is one of the twenty-four countries that carried out round 1 of the Survey of Adult Skills (PIAAC). Adults between 23-65 years old, so it spreads from around the 1950’s until now, participated in this survey and answered questions about several subjects where, for this research, some were selected from. Five columns were chosen and downloaded in an excel file, where further regressions could be done with. The data of the number of persons used, so outliers were filtered, is exactly 1351 individuals. These persons were all females with one or more children. Both the age of the first born child as well as the age of the mother was known. With these two variables, one variable will be made that is used in the regression. The reason for using this data is because of the reliability of OECD data. Lots of people from different countries participated in the survey. In graph 1, a scatter of the years of education is seen for all the persons in the data set, which clarifies the spread of people who answered the survey. Most people have 7, 11, 14 or 16 years of education. In the Netherlands, these are natural years to graduate from elementary school, high school and higher educational levels. Till 1975, children who are 10 years old and younger had compulsory education. After that, this age has increased by law (Beets, 2004). This is why for only some older participants in the data set their length of education is only 7 years.

(13)

13

Graph 1

The variables that were selected, were the education of the mother, which was measured in years of education, the age of the first born child as well as the mother’s in years and the career of the mother which was nominally measured in experience from one till four, where four was no experience and one was currently working. This can be seen in table 1.

Table 1

Composition of nominal experience data Currently working (paid or unpaid) 1 Recent work experience in last 12 months 2 Left paid work longer than 12 months ago 3

No work experience 4

Status unknown 5

- Estimations

In this section, an Ordinary Least Squares (OLS) estimation has been done to estimate the effects of the age of the mother when she gets her first child on her education and on her labour market results, also called her career. The result is a linear effect. This shows the outcome of giving birth to a child one year later compared to the gain in education or experience in the labour market. After this, the data will be

(14)

14 sorted from older to younger people and with this it can be measured whether this described effect has become larger or smaller from the 1950’s until now.

Probably, various other variables will have an effect on the education of the mother, which is the dependent variable in this case. For instance, in many other studies, researchers examined the effect of the parent’s education level on the schooling of the children. This could thus also have an effect on the participants in this study (Holmlund et al., 2011). And, maybe the most important influence on the education of the mother is her own intelligence. This variable or something comparable can unfortunately not be found in the data set. To enlarge the credibility of the model, some variables will be added to the regression. The one variable that is added in the third regression (see table 3) is called “learning at work”; this measures the extent to which a person learns from doing the job she is carrying out. In table 2 the measurement of this variable is seen. This will probably have a negative influence on her education, because she can choose to either skip schooling and thus have less education and instead learn on the job. All other influencing factors will be recorded in the error term and this will cause omitted variable bias. Whether this will affect the estimation positively or negatively cannot be said, because there are so many other influences that have effect on the dependent variable. Some of them have a positive and some a negative influence, and thus nothing can be said about the overall sign of the error term. The one assumption made on the measured data is that there was no unexpected pregnancy. In this way, we assume that the mothers knowingly choose to become pregnant and get a child.

Table 2

Composition of Learning at Work data on scale

1 _{Lowest to 20%}

2 _{More than 20% to 40%}

5 More than 80%

This thesis will do research on two effects of the age of the mother when she gets the first child. Hereby, the first and second hypotheses can be confirmed or not. For the third hypothesis, the data will be sorted in age and grouped into 10 years groups. With this, the estimations will be done again for smaller groups and the

(15)

15 outcomes will be compared. Comparing this, it can be said whether the effect became stronger or weaker through the years. The third hypothesis can be stated because of the size of the data. With the information of 1351 people and 4 groups of 10 years (23-35 years, 36-45 years, 46-55 years and 56-65 years), each group will be large enough to do a reliable estimation.

The hypotheses described above can also be entered up in the following way:

H0 : b = 0 H1 : b¹ 0

The first estimation will describe the relation between the age of the mother when she gets her first child (Ai) on her own education (Ei). This is specified in the following equation (1):

Ei = a + bAi + ui (1)

The second estimation will describe the relation between the age of the mother when she gets her first child (Ai) on her labour market performance, so career (Ci). Equation (2) describes the relation:

Ci = a + bAi+ ui (2)

In the third estimation, the relationship is described between the dependent variable years of education (Ei) and the two independent variables age when the mother gets her first child (Ai) and learning at work (Li). Equation (3) shows the relation:

Ei = a + b1Ai + b2Li + ui (3)

With the outcomes of the regression, a T-test can be done to test the significance. If the resulting T-values are above the value 2, we can declare that the outcome is significant. A significant T-value means that the nil hypothesis will be rejected. The coefficients that will result will describe how strong the relation is between the dependent and independent variables. Furthermore, it will induce if there is negative or positive relation.

(16)

16 - Modifications of the data

Prior to carrying out the regression, some modification of the data was performed. Firstly, for some individuals, specific data was not known in the data set or was filled in a way that could not be an answer to a question. These individuals were filtered out. Secondly, because the representation of the variable ‘career’ was measured in experience and in nominal data, this can bias the outcomes. The regression will simply give an outcome in numbers. For example, if the status of a person was unknown, this person got a 5. This will thus bias the overall effect and that is why persons from whom the status was unknown, were filtered out. Furthermore, just women with at least one child were picked out of the data set.

IV. Results

After doing the analysis on the data with the variables mentioned above, the results will be interpreted with the OLS regression. The first regression done aims at clarifying the effect of the age of the mother when she gets her first born child on her education. Looking at the P-value, the effect of the independent variable on the dependent variable measured equals significantly 0,195 with 1%. This means that if the mother gets her child one year later, her education will increase by 0,195 years. This will have various effects on, for example, her future career. This positive result confirms the hypothesis stated. Explanations for this can be, as said in previous sections, that the mother has more time to spend on her education during the years 16 till 25. However, this linear coefficient only gives an increase of the overall effect for women between 16 and 65 years old. It would be logical if there is not a linear, but a positive exponential effect, which is decreasing. The effect of getting a child on the education of the mother will in her early years, so between the age of 16-30 years, have a much bigger effect compared to the same effect after her 30th year of life. In the literature review, Cools et al. (2017) investigated the short and long term effect of children on labour outcomes of the parents. Their results showed that there is no long term effect of having children on the labour outcome and this could also be the case on education of the mother. Therefore, more research has to be done. Having said this, we can conclude that there is indeed a positive effect of the mother’s age when she gets her first child on her own education and we can confirm hypothesis 1.

(17)

17 The second hypothesis, can also be confirmed with the data that is used. The coefficient that results from the regression with dependent value age of the mother when she gets her first child and independent value the experience now, referring to career, is -0,04. This indicates that if the mother gets her child one year later, her ratio in the given table 1 above will go down with 0,04. This thus shows a small positive effect on her career, because the ratios that are used within this variable have a reverse effect. Therefore, we can conclude that there is indeed a positive effect of the mother’s age when she gets her first child on her future career and we can confirm hypothesis 2.

In table 3, an overview is given of all the coefficients that were regressed. In the third regression, a variable called ‘Learning at work’ was added to increase the R-squared. This variable explains the extent to which a person is educated at work and has thus an effect on education of the participant. The variable ‘years of education’ can have influence, because a person can choose to, for example, have no educational years at all and get educated at work. These years will not be recorded in ‘years of education’.

The third hypothesis is tested in table 4. The resulting coefficients are not as expected. Hypothesis 3 stated that the effect of our independent variable on the dependent variable would increase in time, because feminism became stronger and stronger from the 1950’s until now and would thus cause an increase of women on the labour market, which would make the effect stronger. However, this is entirely not the case as can be seen in the table. It can only be concluded that the effect on her education of the mother’s age when she gets her first child has barely changed through the years. The only remarkable thing which actually satisfies the expectation, is that the intersection value is slightly increasing through the years. This is caused by the fact that schooling became mandatory for more years through the 1970’s and 1980’s.

(18)

18

Table 3

Estimation of the effect of having children

Dependent variable: education and working

experience of the mother

Variables effect on years of education and working experience of the mother

Age when first child was born 0,19489* -0,04108 0,16427*

(0,0134553) _(0,0050352) _(0,0148531) Learning at work 0,35698* (0,0511667) N 1352 1352 895

Representation of the coefficient of estimates on different regressions * 1-percent significance

(19)

19

Table 4

Compared effect of mother's age when she gets her first

child on her education in time

Age group in years Coefficient N 23-35 -0,010141 131 (0,04151177) 36-45 0,0241784 377 (0,02550437) 46-55 -0,0080731 451 (0,02458002) 56-65 0,02452267 391 (0,02842622)

(20)

20

V. Discussion

At first, there is no doubt about the fact that the results show that there is an effect of the age when the mother gets her first child on her education and career.

But after having done this research, a couple of things can be improved to make the effect more credible. First, whether there was more information about the intellectual capabilities of the participants and other information with respect to the education of the subject. If this could be added to the regression, the proportion of explanation of the dependent variable, R-squared would be higher.

There occurs also a problem with the difference in people’s nature and nurture. These separate subjects each have, apart from each other, an influence on the education. This could cause the omitted variable bias which is mentioned above. A solution to this would be to use twins of adoptees in the regression, who have the same nurture, so this unwanted influence is filtered out (Butcher & Case, 1994).

That is where this variable bias thus comes in. The size of the R-squared is low. Variable bias happens to be omitted. Probably, the independent variable used does not have the most impact on years of education, which causes the bias. This can be solved by adding additional independent variables, which is tried in the third regression (see table 3), by adding the variable ‘learning at work’. In the paper of Butcher and Case (1994), they also talk about the influence of sibling sex composition on woman’s education. The advantage of adding variables is that the R-squared increases, but it also has its disadvantages, namely that you can see that variables become less significant. In the third regression in table 3, this is not yet a problem, but if more variables would be added, multicollinearity would appear, which causes insignificance. A solution to this problem would possibly be an IV estimation.

Secondly, in testing the third hypothesis, the number of participants (N) of the group between 23 and 35 years is low. This can cause the insignificance of the resulting coefficient. However, insignificancy is a problem in all the groups. Solution for this problem would thus be to use an enlarged data set.

(21)

21

VI. Conclusion

This research study examined the effect of a mother’s age when she gets her first child on her own education, when using Dutch data. In the last decades feminism became stronger and women were more involved in the labour market. The consequence of this is that the effect of having children also became stronger on both parents. In most of the times, the care of the child rebounds upon the woman in the family, so it is expected that the effect of having children will be stronger on women than on men.

Previous literature has shown that there is indeed a strong overall effect on parents of having children. Angrist and Evans (1998) did a study on the effect of getting children at a younger age on labour market participation for low and high educated parents. The costs that come with taking care of a child are not compensated by either the father nor the mother. Angrist and Lavy (1996) concluded that childbearing at a young age does not necessarily have a really negative effect on the mother’s education and career; the only factor that is important is whether there is presence of the father or not, who can take over some care of the children from her. Cools et al. (2017) did research in addition to Angrist and Evans (1998) and concluded that the examined effect is there in the short and medium term, but that this effect will disappear in the long term, because parents adjust their labour market behaviour in this long term, which mutes the effect.

OECD data has been used to do an estimation of the described effect. Dutch people participated in a survey and answered questions about diverse topics (OECD, 2012).

Results show that most of the participants in the data have 11 to 14 years of schooling. Several OLS regressions have been done with as dependent variables the education of the mother and career of the mother and as independent variable the age of the mother when she gets her first child. This latter independent variable by using the age of the mother and the age of the first born child, is information from the data set. The outcomes of the regression showed that there is a small effect of the independent on the dependent variable. The overall positive linear effect on years of education would go up with almost 20% of a year if the mother chooses to bear her child one year later. Also, the effect of the career of the mother worked out positive. It was measured in a nominal scale which gave a negative coefficient, but this means a

(22)

22 positive effect of the age of the mother when she gets her first child on her career, and, in this case it was measured in experience. Explanations for this are that if this effect would be compared with data from other countries, the Dutch education system and labour market is quite open to women (Beets, 2004). This means that if a woman gets her first child early in her life, she will probably miss out more of her chances on education and the labour market, compared to other countries. On the other side, as said as explanation for the expectations, the Netherlands is a fairly liberal country where everyone has equal chances. However, this will not have an effect on women with children, because in the Netherlands women choose by their self to have children and as a consequence set their education and career aside in most of the times.

Conclusion is that this effect is a negative occurrence in society, because in most cases, there appears to be a backlog for women on their education but also on the labour market compared to men. Solutions for this issue would be policy implications that can mitigate the consequences of becoming a mother. An example of these solutions is to improve the quality of the childcare facilities, because the cause that parents decide not to send their child to such a childcare facility is because the quality is in many cases very low. The quality of childcare facilities could be improved through e.g. government subsidies.

(23)

23 References

Alderman, E. M. (2001, 19 December). The Effect of Children on Parents. Retrieved from

https://jamanetwork.com.proxy.uba.uva.nl:2443/journals/jama/fullarticle/18443 98

Angrist, J. D., & Evans, W. N. (1998, June). Children and Their Parents' Labor Supply: Evidence from Exogenous Variation in Family Size. Retrieved from

https://www-jstor-org.proxy.uba.uva.nl:2443/stable/pdf/116844.pdf?refreqid=excelsior%3Aa1f5b d1fc7ce1f242da27e15e9197ed0

Angrist, J. D., & Lavy, V. (1996, October). The effect of teen childbearing and single parenthood on childhood disabilities and progress in school. Retrieved from http://www.nber.org.proxy.uba.uva.nl:2048/papers/w5807.pdf

Balen, F. (2004). Late parenthood among subfertile and fertile couples: motivations and educational goals. Retrieved from

https://ac.els- cdn.com/S0738399104002782/1-s2.0-S0738399104002782-

main.pdf?_tid=0c8ee162-f17c-452f-979c-e590e3aff822&acdnat=1523280719_a5b7cedfe4ee44ec91de83ff984d3041 Beets, G. (2004). De timing van het eerste kind: een overzicht. Retrieved from

https://www.nidi.nl/shared/content/output/2004/bg-33-01-beets.pdf

Butcher, K. F., & Case, A. (1994, August). The Effect of Sibling Sex Composition on Women's Education and Earnings. Retrieved from

https://www-jstor-org.proxy.uba.uva.nl:2443/stable/pdf/2118413.pdf?refreqid=excelsior%3A1c0 4d89fe0abcc9652ab8359fb8d6959

Cools, S., Markussen, S., & Strøm, M. (2017, 6 September). Children and Careers: How Family Size Affects Parents’ Labor Market Outcomes in the Long Run. Retrieved from https://link.springer.com/content/pdf/10.1007%2Fs13524-017-0612-0.pdf

Cruses, G., & Galiani, S. (2005, 18 November). Fertility and female labor supply in Latin America: New causal evidence. Retrieved from https://ac-els-cdn-

com.proxy.uba.uva.nl:2443/S0927537105000771/1-s2.0-

(24)

24 Finley, G. E. (1998). Parental Age and Parenting Quality as Perceived by Late

Adolescents. Retrieved from

https://www-tandfonline-com.proxy.uba.uva.nl:2443/doi/pdf/10.1080/00221329809596167?needAcces s=true

Holmlund, H., Lindahl, M., & Plug, E. (2011, September). The Causal Effect of Parents’ Schooling on Children’s Schooling: A Comparison of Estimation Methods. Retrieved from

https://pubs-aeaweb-org.proxy.uba.uva.nl:2443/doi/pdfplus/10.1257/jel.49.3.615

Morris, M. (1989, March). Last-Chance Children: Growing Up With Older Parents. Retrieved from

http://www.jstor.org.proxy.uba.uva.nl:2048/stable/pdf/2074137.pdf?refreqid=excelsior %3A5e065356bb8e686058a1cb80f7273339

OECD. (z.d.). Survey of Adult Skills (PIAAC) [Dataset]. Retrieved from http://www.oecd.org/skills/piaac/publicdataandanalysis/#d.en.408927

Van Noord-Zaadstra, B. M., Looman, C. W. N., Alsbach, H., Habbema, J. D. F., Te Velde, E. R., & Karbaat, J. (1991, 8 June). Delaying childbearing: effect of age on fecundity and outcome of pregnancy. Retrieved from https://www-bmj-com.proxy.uba.uva.nl:2443/content/bmj/302/6789/1361.full.pdf

(25)

25

Appendix 1: OLS Estimates

These estimates are done in excel data analysis.

Table 5.1

SUMMARY

Dependent variable: years of schooling

Information for the regression

Multiple correlationcoefficient R 0,366750003 R-squared 0,134505565 Adjusted R-squared 0,133864458 Standard error 2,433609135 Observations 1352 Variance-analysis

Degrees of freedom squares Sum of Average sum of squares F Significance F

Regression 1 1242,542908 1242,542908 209,8020566 2,65258E-44

Error 1350 7995,312121 5,922453423

Total 1351 9237,85503

Coefficients Standard error T- statistical information P-value Lowest 95% Highest 95%

Intersection 7,804204622 0,364064805 21,43630616 5,67762E-88 7,090010403 8,51839884

(26)

26

Table 5.2

SUMMARY _{Dependent variable: career mother}

Multiple correlationcoefficient R 0,216825391 R-squared 0,04701325 Adjusted R-squared 0,046306811 Standard error 0,910285929 Observations 1351 Variance-analysis Degrees of

freedom Sum of squares Average sum of squares F Significance F

Regression 1 55,14435018 55,14435018 66,54958693 7,73648E-16

Error 1349 1117,809018 0,828620473

Total 1350 1172,953368

Coefficients Standard error T- statistical information P-value Lowest 95%

Highest 95%

Intersection 2,652304794 0,136218136 19,47101084 1,30178E-74 2,385082398 2,91952719

(27)

27

Table 5.3

SUMMARY Dependent variable: years of schooling

Multiple correlationcoefficient R 0,4174644 R-squared 0,174276525 Adjusted R-squared 0,172425127 Standard error 2,14480828 Observations 895 Variance-analysis Degrees of

freedom Sum of squares

Average sum of squares F Significance F Regression 2 866,056191 433,0280955 94,13239747 8,09554E-38 Error 892 4103,380681 4,600202557 Total 894 4969,436872

Intersection 8,047621131 0,419171316 19,19888319 4,99082E-69 7,224944175 8,870298087

Age when first child was born 0,164267511 0,014853063 11,05950417 9,83105E-27 0,135116489 0,193418534

(28)

28

Table 6.1

SUMMARY

Participants 23-35 years old

Multiple correlationcoefficient R 0,02150369 R-squared 0,000462409 Adjusted R-squared -0,007285945 Standard error 2,248911873 Observations 131 Variance-analysis

Regression 1 0,301829317 0,301829317 0,059678314 0,807393226

Error 129 652,4309951 5,057604613

Total 130 652,7328244

Intersection 13,76057532 1,130732192 12,16961489 3,6478E-23 11,52339399 15,99775666

Age when first child was born -0,01014097 0,041511766 -0,244291453 0,807393226

(29)

29

Table 6.2

SUMMARY

Regression 1 5,673009456 5,673009456 0,89872318 0,343734809

Error 375 2367,112136 6,31229903

Total 376 2372,785146

Intersection 12,93125873 0,678143889 19,06860616 3,71679E-57 11,59781751 14,26469995

(30)

30

Table 6.3

SUMMARY

Regression 1 0,723643331 0,723643331 0,107873819 0,742730987

Error 449 3011,999195 6,708238741

Total 450 3012,722838

Intersection 13,26972094 0,670378741 19,79436419 3,75081E-63 11,95225143 14,58719045

Age when first child was born -0,0080731 0,024580024 -0,3284415 0,742730987

(31)

31

Table 6.4

SUMMARY

Regression 1 5,313807173 5,313807173 0,74421316 0,388843396

Error 391 2791,805786 7,14016825

Total 392 2797,119593

Snijpunt 11,54439675 0,772985581 14,93481514 3,16566E-40 10,0246687 13,0641248

(32)

32

Appendix 2: Expected relationship

Graph 2

Vertical-axis: Years of education (years)

Horizontal-axis: age when the mother gets her first child (years)

0 2 4 6 8 10 12 14 16 18 1 5 9 1 3 1 7 2 1 2 5 2 9 3 3 3 7 4 1 4 5

RELATIONSHIP AGE WHEN THE MOTHER GETS HER FIRST

CHILD AND YEARS OF EDUCATION

(33)

33

Graph 3

Vertical-axis: Career (measured in nominal experience data) Horizontal-axis: Age when the mother gets her first child (years)

0 0,5 1 1,5 2 2,5 3 1 6 1 1 1 6 2 1 2 6 3 1 3 6 4 1 4 6 5 1 5 6 6 1

Age effects on mothers’ education with respect to the first born child

Thesis

Name: Coen van den Brink

Student number: 10775471

Specialization: Economics

Field: Labour Economics

Number of credits: 12 EC

Title: Age effects on mothers’ education with respect to the first

born child

Appendix 1: OLS Estimates

Appendix 2: Expected relationship

RELATIONSHIP AGE WHEN THE MOTHER GETS HER FIRST

CHILD AND YEARS OF EDUCATION

RELATIONSHIP AGE WHEN MOTHER GETS HER FIRST

CHILD AND CAREER