The Motherhood penalty for Dutch women : a comparison between the Netherlands and other European countries

(1)

The Motherhood penalty for Dutch women: A

comparison between the Netherlands and other

European countries

Bachelor thesis

Cheyenne Schouws

Student number: 10975462

Supervisor: Melvin Vooren

June 26

th

₂₀₁₈

University of Amsterdam

Faculty of Economics and Business

BSc Economics and Business

Specialisation: Economics

(2)

Statement of originality

Hereby I, Cheyenne Schouws, declare that I have written this thesis myself and that I take full responsibility for its content. I confirm that the work presented in this thesis is original and that no sources other than those mentioned in the text and in the references are used. The Faculty of Economics and Business (FEB) is only responsible for the supervision to complete this thesis, not for the content.

(3)

Abstract

This paper will provide a descriptive analysis of the effect that children have of on women’s wages for eleven European countries, a so-called motherhood wage penalty. In several European countries this penalty exists, whereas for some countries mothers do not have any disadvantage. The focus is mainly on the relative situation of the Netherlands, for which the effect of children on female wages has not been investigated yet. The gender wage gap of the Netherlands is larger for mothers than for childless women. This might indicate the existence of a motherhood wage penalty for Dutch women. By using data from the OECD’s Survey of Adult Skills, OLS-regressions are done. The results show that there is no significant effect of having children on the hourly wage of Dutch women, however Dutch mothers are relatively in a better situation compared to mothers from the United Kingdom.

(5)

1. Introduction

According to Statistics Netherlands (CBS) the gender wage gap in the Netherlands is largest for employees with children. Dutch women who work in the private sector earn on average 14 percent less than men do for the same job. This difference in wage is three times smaller for employees without children. For the public sector the gender wage gap for employees with children is 5.8 percent, whereas the wage gap for employees without children is only 2.4 percent (CBS, 2014).

The wage gap between men and women appears to be largest in the age group between 25 to 35 years. This is exactly the age in which many women have young children. Children are especially damaging to the career high-educated women with high paid jobs (Miller, 2017). Studies have shown that both in the US and in Europe the gender pay gap is much smaller until the first child is born. Even though women earn less than men do, earnings of women without children keep growing at a similar rate to men whereas the income of women with children drops sharply (Miller, 2018). A Danish study by Kleven, Landais and Søgaard showed that this drop was about 30 percent, in the long run mothers earn 20 percent less. According to this study, having children accounted for 80 percent of the gender wage gap in 2013.

According to CBS (2016), there does exist a gender wage gap in the Netherlands, but this is for a large part accountable to characteristics like schooling or job type (full- or part-time work).

It is often seen that Dutch women start to work part-time after having children, rather than completely dropping out of the labour market (Eurostat, 2016). Besides, CBS claims that this gap is getting smaller because of the increasing educational level of women (CBS, 2016).

Even though the gender wage gap is narrowing, there is still a larger gap for employees with children than without children. Is this due to a motherhood wage penalty for Dutch women? Since the educational level of Dutch women has increased over the past decade, the penalty on motherhood might be relatively large. The answer to this question may give insights in the (financial) consequences that a woman will face when she will have her first child. Besides, the answer might contribute to the already existing evidence on the unequal consequences of having children between men and women.

There has already been done a lot of research on the effect of having children on women’s earnings. However, the effect of having children on the income of Dutch women specifically has not yet been investigated. If Dutch women have a motherhood wage penalty, how bad is the situation for Dutch women in comparison to women from other European countries? Therefore, the research question of this paper will be: What is the effect of having

(6)

children on the earnings of Dutch women and how is the situation of the Netherlands in comparison to other European countries?

This paper will provide a descriptive analysis on the existence of a negative effect of having children on women’s earnings, a motherhood wage penalty. In addition it will describe how the Netherlands is doing compared to other European countries. To find an answer, two models are derived; one that includes all chosen countries and one model with dummy variables for every country and country-specific interaction variables with motherhood. To find a reliable effect, several control variables are added. With the use of these models, OLS-regressions will be run. These regressions will be done with data from the Survey of Adult Skills (PIAAC) of 2012. A data set that contains information on demographic and labour market characteristics of many different countries. The results of both regressions show, that there is no motherhood wage penalty for the Netherlands. Looking at all countries included, only one of them has a significant motherhood wage penalty. Actually, for most countries included having children has no significant effect on women’s earnings.

The remainder of this paper will be structured as followed. Section two consists of a literature review to familiarise with this topic and to give insights in previously done research and their findings. Next, a description and motivation of the data that is used will be provided followed by the methodology. Section five will contain the results. After this, a conclusion will be drawn and discussed. Finally, the limitation of this research will be explained in section seven.

(7)

2. Literature review

Over the past decades, women have increased their years of education substantially relative to men. They obtained more job experience and female labour force participation increased. Many women had the desire to attain a good career and wanted to take good care of their families at the same time; especially college-graduated women had this aim. Women started to work more hours a week and female earnings increased relatively to that of men (Goldin, 2014).

In the seventies and eighties, the gender wage gap decreased considerably in many countries. Despite the increased amount of human capital women had accumulated, narrowing of the gender wage gap stagnated. A possible explanation for this is the fact that many women need to put their career on hold, mainly because of their pregnancy and child rearing. Another factor why narrowing of the gender wage gap stagnated, might be the fact that many counties extended maternal and parental leave schemes. These schemes are commonly more often used by women than by men. This may have caused women to stay out of the labor market for longer periods (Gupta & Smith, 2007).

According to traditional human capital theory (Mincer & Polachek, 1974), the motherhood wage penalty can be explained by three reasons. First, as soon as a woman takes time off for child rearing reasons, she stops accumulating human capital. Second, if existing human capital is not used for a while, depreciation of human capital can occur. Finally, firms who pay a lot for on-the-job training may not be willing to assign such jobs to women. The employer might expect women to have children one day; therefore they might need to use maternity leave. Thus, men have higher efficiency, which makes investing in women less attractive (Gupta & Smith, 2007).

One day a mother will return to the labor market, then two things can happen. She easily catches up on her lost human capital, and then she will have a steep wage profile. On the other hand, there is a high possibility that a mother will return to a job that has fewer promotion opportunities, and then she will not have striking career prospects. In addition, it can be the case that women self-select jobs with less good career prospects before they even have children (Gupta & Smith, 2007).

For a large part, the existence of the gender wage gap has always been assigned to human capital differences. Nowadays these differences are much smaller but still a large wage gap between women without children and mothers can be distinguished. According to Goldin (2014), children are the main reason why a woman has to leave the labour force for a while or why she would make a labour supply change. Women with children work on average 24 percent fewer hours a week than men or women without children with comparable characteristics would do.

(8)

Despite the narrowing of the gender wage gap, women still earn less than men. According to Waldfogel (1997) not only a gender wage gap exists, but also a “family gap”. She mentions that this family gap is the difference in hourly wages between women with and without children. According to Human capital theory, Becker (1985) claims that women spend more time outside the labour force when having a child. The labour market experience that these women will miss, explains a large part of the family gap. In contrast, Hill (1979) found that the family gap narrows almost completely as soon as labour market experience is being controlled for. However, in 1994, Waldfogel herself found that controlling for labour market experience could only explain the family gap partially. Still, unexplained effects of having children remained.

Waldfogel considered two possible explanations for the unexplained difference in wage between mothers and childless women. The first possible explanation is unobserved heterogeneity, this means unobserved differences in characteristics. Korenman and Neumark (1992) found evidence on unobserved heterogeneity, however two years later by, using another method, they did not find a heterogeneity bias. Waldfogel’s next possible cause is part-time employment. She substantiates this with evidence from Blank (1990). Part-time jobs could lead to lower wages because in that case, less human capital gets accumulated. Besides, there is a higher probability that employers do not allocate wage raises to workers with a part-time job. Waldfogel (1997) concludes that the motherhood wage penalty does not disappear even when controlling for labour market experience and marital status. Leaving the labour market for a while, because of childbirth, often leads to lower wages. The wage penalty she found for having one child is 6 percent and this penalty reaches 13 percent for having two children or more. She found no evidence that unobserved heterogeneity contributes to the family gap. Current part-time employment, on the other hand, does contribute to the family gap.

In line with Waldfogels findings, Budig and England (2001) also found evidence of a motherhood penalty for American women. One explanation for the motherhood penalty Budig and England share with Waldfogel (1997), is the fact that women lose job experience by interrupting their career; they leave the labour force for a while to take care of their children or start working part-time. Even though nowadays 40 percent of all mothers with children under the age of one are employed, every mother loses some job experience at one point (Klerman & Leibowitz, 1999). Even when women leave the labour market for a short period, this will cause a loss of work experience. According to human capital theory, experience and thus seniority will yield positive returns; therefore, missing job experience will cause wages to stagnate at a certain level. In addition, less job experience signals lower productivity.

According to Becker (1991), women with children are (expected to be) less productive in doing their job than women without children. Mothers may be tired from child rearing or they

(9)

try to save some energy for their children at home, which lowers productivity. Women without children will have much more free time, which means much more time to sit back. Budig and England (2001) mention lower productivity as a second explanation why mothers earn less. Moreover, Budig and England claim that women may decide to sacrifice their high wages and switch to a “mother-friendly” job that fits parenting. This is comparable with self-selection into certain jobs, where Gupta and Smith (2007) wrote about. Because specific women (mothers) are looking for these mother-friendly jobs, employers can set lower wages for jobs with mother-friendly characteristics. Mothers are willing to trade off lower wages for mother-friendly characteristics, therefore they will earn less. Budig and England found a motherhood wage penalty of 7 percent. One third could be explained by job experience and seniority, and two-thirds remained undefined. They found that having a job with mother-friendly characteristics could only explain the motherhood penalty for a small part. Half of this effect could be accrued to currently having a part-time job. In total no more than one-third of the motherhood penalty could be assigned to career interruption.

In a next study, Waldfogel (1998) tries to give insights into why the "family gap" might exist. In earlier research, she found that having children has a negative effect on women’s earnings. In the United States, it is obvious that women with children earn less than women without children, even after controlling for differences in characteristics like education and labour market experience. Besides mothers are less likely to have a successful career (Goldin, 1990). This penalty does not exist for men.

Over the past years, women increased investment in education, this caused the gender wage gap to decrease. Even though the wage gap between men and women is narrowing, the family gap is increasing. Waldfogel claimed, that because of the increasing family gap, the gender wage gap in the United States remained larger than for other countries. A possible explanation for this is the fact that the United States provides bad access to job-protected maternity leave. This creates a barrier for mothers to enter jobs that acquire work experience (Waldfogel, 1998). This new hypothesis of Waldfogel (1998) on maternity leave is supported by research of Jacobsen and Levin (1995), who found that breaks in employment due to motherhood have a lifelong effect on women’s wages.

Rosenfeld and Kalleberg (1991) came up with a complementary explanation why “bad” maternal leave schemes might lead to lower female wages. They wrote about statistical discrimination. It is expected that practically every mother will make use of maternal leave schemes. Because of this expectation, employers will automatically assign young women into jobs with less on-the-job training, low promotion profiles and thus smaller possible wage increases (mother-friendly). It is expected that women will have one or two children on average. Therefore, they will take time off once or twice as long as the maternal leave period

(10)

allows.

If the maternal or parental leave schemes are relatively long and the difference of usage of the schemes between man and women is large, statistical discrimination will be large. The theory of Rosenfeld and Kalleberg (1991) is in line with signaling theory. Even though women do not signal low productivity by using a maternal leave scheme, these schemes will have a negative effect on all women, even women who may never have kids. So, as soon as maternal leave schemes become too generous, overall women’s wages will be negatively affected. (Gupta & Smith, 2007).

Similar to these results Waldfogel (1998) found in the United States, a family gap has been found in the United Kingdom. The United Kingdom had the longest maternity leave period. In addition, the United Kingdom had the worst childcare provision and the rate of part-time working mothers was the highest (Joshi, 1991).

In contradiction to studies in the United States and the United Kingdom, for Danish women no motherhood wage gap has been found. Rosholm and Smith (1996) and Naur and Smit (1997), both did not find a motherhood penalty for Danish women. Moreover, Naur and Smit (1997) did find a significant negative effect of being married for women. Being married, even without having kids, tends to lower female income. For males, marriage has a positive effect on earnings.

According to Gupta and Smith (2007), there is no long-term motherhood wage penalty for Danish women. Because of the length of Danish maternal and parental leave schemes, career interruptions do not have a negative effect on the long-term wages of Danish women. Almost every woman takes a substantial amount of time off for motherhood. This results in less difference in the length of staying out of the labour market. Therefore, women do not get the opportunity to signal low commitment and productivity. They did found some negative effect of marriage and having children on women’s wages, however, this disappears when Gupta and Smith (2007) use a panel data estimator that controls for unobserved differences in characteristic and self-selection. In conclusion, motherhood does have a negative effect on women’s wages, but this negative effect disappears after a certain age. Holding experience constant, Gupta and Smith (2007) conclude that having children does not have a significant long-term effect on wages of Danish mothers.

But why should we care about the existence of the motherhood penalty? According to Budig and England (2001), the motherhood penalty plays an important role in gender inequality. A large part of women has children which means a large part of women suffers from a motherhood penalty. Mothers pay a price that does not have to be paid by fathers. Women's earnings are not only lowered for a short period of time, by staying home for a while they will have a lower income for the rest of their life (Davies & Joshi, 1995). Besides, the motherhood

(11)

penalty creates a poverty gap between single mothers and mothers living with a spouse (McLanahan & Kelly, 2006). It is possible for women without children to earn almost exactly as much as men do, while mothers cannot. Single mothers earn even less than mothers with a spouse, single mothers earn only 56-66 percent of what a man earns (Waldfogel, 1998). In addition, Waldfogel mentions that it will be impossible to achieve full equality of pay for women as long as they will be disadvantaged by motherhood. With the existence of a family gap, the gender wage gap will remain. From an economic point of view, disadvantaging mothers will make it impossible to fully utilise human capital of all women.

Buding and England (2001) talk about “free-riding” on mothers’ child rearing. They mention that good parenting provides a child, which will grow up to be a productive adult. This economic productivity will mostly benefit the future employer and other persons surrounding this well-reared individual. None of them will ever pay the parent. This means that a mother pays a price by having a lower wage as a consequence of taking care of her child, while other people will reap the benefits of her child’s labour in the future.

Just as mentioned earlier, also in the Netherlands the labour market participation of women has increased. The difference in productivity between men and women is disappearing. In addition, changes in regulation and antitrust laws have had a positive effect on the gender wage gap (Van der Meer, 2008). The Netherlands still has a gender wage gap of around 20 percent while the average gap in Europe is 17 percent. Only 25-30 percent can be explained by productivity differences. The largest part is explained due to “price differences”, meaning men are slightly paid too much and women are slightly underpaid (Van der Meer, 2008). According to Fransen, Pantenga and Vlasblom (2012) the pay differential can be explained by human capital theory and discrimination.

When looking at evidence of research on the motherhood penalty in other countries, it might be the case that also in the Netherlands a motherhood penalty contributes to the gender wage gap. The “price differences” that Van der Meer (2008) mentions, may include a difference in pay between women with and without children. Fransen et al. (2012) mention that human capital can explain the gender wage gap. This includes work experience, a factor that seems to have a lot to do with the wage penalty on motherhood. It seems possible that also Dutch women are subject to a motherhood wage penalty. The research question is: What is the effect of having children on Dutch women and how is the situation of the Netherlands compared to other European countries? Because of the features of the dataset, first the overall motherhood wage penalty for all countries included will be estimated.

(12)

After that, the isolated effect on Dutch women will be estimated. This leads to the following sub questions:

- Is there an overall motherhood wage penalty for the European countries included? - What is the effect of having children on Dutch women?

- How is the situation of the Netherlands compared to other European countries? To answer these questions the following hypotheses will be tested:

Hypothesis 1 There does exist a motherhood wage penalty for the included European countries.

Hypothesis 2 There does exist a motherhood wage penalty for Dutch women Hypothesis 3 The Netherlands will do better than Greece and the United Kingdom Even though not every researcher has found a negative effect of motherhood, the expectation is that the data will show that there is a motherhood wage penalty for the included European countries, not every European country is as well organised as Denmark. Since the Netherlands still has a gender wage gap, even though women increased their human capital overtime, I think that there is a possibility that a motherhood penalty for Dutch women exists. However, I expect that the Netherlands will be in a relatively better position than some other countries. Several researchers have found, that leaving the labour force for a while will have a lifelong negative effect on women’s earnings. (Waldfogel, 1997; Goldin, 2014; Becker 1985, Jacobsen & Levin, 1995).

Taking time off, because of motherhood, will stop the accumulation of human capital for a while. According to Mincer and Polachek (1974) less human capital will lead to lower income. The length of maternity leave can influence the length of the time out of the labour market. According to Gupta and Smith (2007), when maternity leave schemes are very generous women will lose more job experience.

(13)

Taking all these findings into account, I expect that countries with maternity leave schemes that allow many weeks off, have a larger motherhood wage penalty. Therefore I expect the Netherlands to have a smaller motherhood effect than Greece and the United Kingdom. As shown in table 1 these countries have the most generous amount of weeks off.

Length, in

weeks Average payment rate (%) Full-rate equivalent, in weeks

Belgium 15.0 64.1 9.6 Czech Republic 28.0 62.6 17.5 Denmark 18.0 53.6 9.6 France 16.0 94.2 15.1 Greece 43.0 54.2 23.3 Italy 21.7 80.0 17.4 Netherlands 16.0 100.0 16.0 Norway 13.0 97.9 12.7 Poland 20.0 100.0 20.0 Spain 16.0 100.0 16 United Kingdom 39.0 30.9 12.1

Table 1. Paid maternity leave entitlements available t mothers, 2016. Source OECD (2017). Retrieved from: www.oecd.org/els/soc/PF2_1_Parental_leave_systems.pdf

(14)

3. Data description

To conduct this research paper, the Survey of Adult Skills is used, a product of the Programme for the International Assessment of Adult Competencies (PIAAC), which is provided by the Organisation of Economic Cooperation and Development (OECD). This programme measures adult skills based on three competencies that are important to function in today’s society; literacy, numeracy, and problem-solving capacity in technology-rich environments. The results of this survey are meant to help countries to better understand how education and training systems influence these competencies. In addition, educators, policymakers and labour economists can use this information to provide new policies that will increase adult skills. Despite this paper is not focused on adult skills; the Survey of Adult Skills provides very useful information on individuals that can be used to find out if a motherhood wage penalty exists. The Survey of Adult Skills contains a background questionnaire that collects, among other things, information on demographics, educational attainment and participation, and labour force status and employment. This part of the survey will be used to conduct this research.

Approximately 5000 individuals between the age of 16 and 65 in each country participated in the Survey of Adult Skills. The survey has been done twice, the first set of data was collected in 2011-2012 and involved 24 countries, and the second round was conducted in 2014-2015 and involved nine additional countries. The participating countries are: Australia, Austria, Belgium, Canada, the Czech Republic, Cyprus, Denmark, Estonia, Finland, France, Germany, Ireland, Italy, Japan, Korea, the Netherlands, Norway, Poland, the Russian Federation, the Slovak Republic, Spain, Sweden, the United Kingdom, the United States and for the second round Chile, Greece, Indonesia, Israel, Lithuania, New Zealand, Singapore, Slovenia and Turkey. (OECD, 2018).

Besides the data of the Netherlands (2012), the data of 10 additional European countries will be used. This will be done to be able to make a comparison between the Netherlands and other European countries, which will give insights on how good or bad the situation of the Netherlands actually is. From the first round, Belgium, the Czech Republic, Denmark, France, Italy, Norway, Poland, Spain and the United Kingdom will be used. From the second round, only Greece will be used.

To make the data usable, only a selected set of variables of the survey are chosen. Since the effect of children on earnings will be explored it is useful to include some additional control variables. Besides demographic characteristics like age, gender and family composition (having a spouse and/or children), the following variables have been selected according to human capital theory: level of education, part-time or full-time employment, and seniority. Also, Waldfogel (1997) controlled for these variables, but for two of them in a slightly different way.

(15)

Waldfogel used one-third of the potential work experience (age – years of education – 6) as experience whereas in this paper seniority (years of experience with the respondent’s current job) will be used as experience. In the paper of Waldfogel (1997), education was defined in years whereas in this paper education is defined in International Standard Classification of Education (ISCED) levels of 1997, varying from 1 to 6. Level 6 means that the respondent has achieved the highest attainable educational level (PhD). Within these six levels different sub-categories are specified. Because of these sub-categories, 14 different educational levels could be chosen in the survey. Since the ISCED levels are not linearly related they will be added as dummy variables. Earnings will be defined in hourly wage. Since there are several countries used that have different currencies, hourly wage will be standardised so that the results can be compared. The research will only be done on employed women between the age of 24 and 55 since this category seems most relevant to the research question.

The variables used in this paper are chosen according to the human capital theory of Mincer and Polachek (1974). Mincer and Polachek discussed that the motherhood wage penalty can be explained on the basis of the amount of human capital mothers acquire. Part-time work, short seniority, and low education will yield a lower amount of human capital. Therefore, these variables seem important to take into account.

The focus of this paper will be on the independent variable of parenthood. Parenthood means that an individual has at least has one child. If parenthood has a negative effect on female earnings, after controlling for several demographic and human capital characteristics, a motherhood wage penalty exists. The dependent variable will be the standardised hourly wage.

(16)

4. Methodology

The statistical method that is used in this paper is a linear regression performed with Stata, to provide OLS estimators. The models that will be used are chosen in response to previous research of Waldfogel (1997) and Mincer and Polachek (1974) as mentioned above. To find out if having a child has an effect on female wages, every regression will be done under the “if condition” that the person is female. By adding the countries and interaction terms between

parent and country as dummy variables, the difference in effect per country can be derived.

The independent variable in question is the dummy variable parent and the dependent variable std_wage is the standardised wage on country level. In addition, several control variables will be used. The definition of the variables is shown in table 2. In table 3 the value of the educational levels is shown.

Variable Value

Std_wage Standardised hourly wage (0, 1) 1

Age 24 - 55

Parent 1 = at least one child, 0 = childess Gender 1 = female, 0 = male

Spouse 1 = living with a spouse, 0 = no spouse Job status 1 = part-time, 0 = full-time

Seniority # Years

Education Level 2-14(dummy) 2 Country Country name (dummy)3

1_{with mean 0 and standard deviation 1, standardised on country level} 2_{Level 1 as reference level. For value of educational levels see table 3} 3_{The Netherlands as reference country, (n-1) country dummies included}

Table 2. Variable list with definition

Level Education Code ISCED level

1 Pre-primary, primary 1 No formal qualification 2 ISCED 1

2 Lower secondary 3 ISCED 2

3 Upper secondary 4 ISCED 3C (< 2y) 5 ISCED 3C (2y >) 6 ISCED 3A-B

7 ISCED 3 without distinction, 2y+ 4 Post-secondary non-tertiary 8 ISCED 4C

9 ISCED 4A-B

10 ISCED 4 without distinction 5 First stage tertiary 11 ISCED 5B

12 ISCED 5A, bachelor degree 13 ISCED 5A, master degree 6 Second stage tertiary 14 ISCED 6, PhD

1_{For specific details on ISCED levels per country see Classifying Educational Programmes,} OECD, 1999. http://www.oecd.org/education/1841854.pdf

(17)

Education will consist of 13 dummy variables, which is a lot to show in a model. Therefore the 13 education dummy variables will be displayed as ∑ 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑖𝑖_{. The corresponding}

coefficients will be displayed as 𝛽𝛽𝑖𝑖. This also holds for the 10 country dummies. These will be

displayed as ∑ 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑐𝑐𝑐𝑐𝑥𝑥_{with corresponding coefficient 𝛽𝛽}

𝑥𝑥. To find out if there exists a

motherhood wage penalty for the included European countries, the following model will be used under the IF condition female.

𝑆𝑆𝑒𝑒𝑒𝑒𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤= 𝛽𝛽0+ 𝛽𝛽𝑤𝑤𝑤𝑤𝑤𝑤 𝑒𝑒𝑎𝑎𝑒𝑒 + 𝛽𝛽𝑗𝑗𝑗𝑗𝑗𝑗 𝑗𝑗𝑒𝑒𝑗𝑗 𝑠𝑠𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑠𝑠 + 𝛽𝛽𝑠𝑠𝑤𝑤𝑠𝑠𝑖𝑖𝑗𝑗𝑠𝑠𝑖𝑖𝑠𝑠𝑠𝑠 𝑠𝑠𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑐𝑐𝑒𝑒𝑒𝑒𝑐𝑐 + 𝛽𝛽𝑠𝑠𝑠𝑠𝑗𝑗𝑠𝑠𝑠𝑠𝑤𝑤 𝑠𝑠𝑠𝑠𝑒𝑒𝑒𝑒𝑠𝑠𝑒𝑒

+ 𝛽𝛽𝑠𝑠𝑤𝑤𝑠𝑠𝑤𝑤𝑠𝑠𝑠𝑠 𝑠𝑠𝑒𝑒𝑐𝑐𝑒𝑒𝑒𝑒𝑒𝑒 + 𝛽𝛽𝑖𝑖� 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑖𝑖 + 𝛽𝛽𝑥𝑥� 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑐𝑐𝑐𝑐𝑥𝑥

In order to explore what the penalty is for every single country involved and whether the Netherlands is doing better or worse,an interaction term of parent and country will be added to the model. Since there will be 10 interaction dummies included, the interaction terms will be presented in the model as ∑ 𝑠𝑠𝑒𝑒𝑐𝑐𝑒𝑒𝑒𝑒𝑒𝑒 ∗ 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑐𝑐𝑐𝑐𝑥𝑥_{with the corresponding coefficient 𝛽𝛽}

𝑗𝑗. The

model will look as following:

𝑆𝑆𝑒𝑒𝑒𝑒

𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤

= 𝛽𝛽

0

+ 𝛽𝛽

𝑤𝑤𝑤𝑤𝑤𝑤

𝑒𝑒𝑎𝑎𝑒𝑒 + 𝛽𝛽

𝑗𝑗𝑗𝑗𝑗𝑗

𝑗𝑗𝑒𝑒𝑗𝑗 𝑠𝑠𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑠𝑠 + 𝛽𝛽

𝑠𝑠𝑤𝑤𝑠𝑠𝑖𝑖𝑗𝑗𝑠𝑠𝑖𝑖𝑠𝑠𝑠𝑠

𝑠𝑠𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑐𝑐𝑒𝑒𝑒𝑒𝑐𝑐 + 𝛽𝛽

𝑠𝑠𝑠𝑠𝑗𝑗𝑠𝑠𝑠𝑠𝑤𝑤

𝑠𝑠𝑠𝑠𝑒𝑒𝑒𝑒𝑠𝑠𝑒𝑒 +

(18)

5. Results

In this chapter, the results of the regression analysis will be presented. Table 4 shows the results of three different regression models. Model (1) is provided to see the effect of parenthood on women, without controlling for demographic, educational, labour characteristics and country. Model (2) is run with the use of these control variables. Finally, model (3) is run with additional interaction terms to isolate the effect of parenthood on women of a specific country. To show if there is any correlation between parenthood and the added control variables a correlation matrix will be provided.

The dependent variable std_wage is the standardised hourly wage of women on country level. Therefore, the results will be interpreted in standard deviation. As can be seen from the results in model (1) there is no evidence for the existence of a motherhood wage penalty, without the use of control variables. Actually, there is a statistically significant effect of 0.0653 standard deviation. This indicates that there is a slight wage advantage for women with children. When control variables are included in model (2), there is still a statistically significant positive standard deviation of 0,0493. This means that hypothesis 1 does not hold. There is no evidence for the existence of an overall motherhood wage penalty for the women of the 11 countries included in this analysis.

According to the result of model (3), there does not exist a motherhood penalty for Dutch women. The Netherlands is the reference country in this model, so the coefficient 0.0536 for

parent says something about the effect of having children on Dutch women. However this effect

is not significant. Therefore there is no evidence for the existence of a motherhood wage penalty for Dutch women. Regarding this result, hypothesis 2 does not hold. Since the Netherlands is the reference country, the effect of having children in the other countries can be derived from the interaction terms parent*country. These findings will be relative to the situation of the Netherlands.

As can be seen in table 4, only for Greece and the United Kingdom the interaction term is significant. This means these countries are the only two countries that differ statistically from the Netherlands. The interaction variable for the United Kingdom shows a significant effect of −0.166 standard deviation, which means the net effect of having children in the United Kingdom, equals0.0536 − 0.166 = −0.1124. For Greece the interaction term has a 0.297 standard deviation, which means that the net effect of having children on Greek women is

(19)

OLS Estimates – motherhood wage penalty

Standardised

(1)

(2)

(3)

wage

Parent

0.0653***

0.0493*

0.0536

(3.41)

(2.37)

(0.88)

Age

0.00695***

0.00662***

(6.29)

(5.94)

Job status

-0.0239

-0.0246

(-1.34)

(-1.36)

Seniority

0.00906***

0.00898***

(8.34)

(8.28)

Spouse

0.0608**

0.0565**

(3.11)

(2.85)

Educational level

3

_{Yes Yes}

Country

4

_{Yes Yes}

Parent*Country

5

1 Belgium

0.0972

(1.10)

2 Czech Republic

-0.0668

(-0.74)

3 Denmark

-0.0990

(-1.13)

4 France

-0.0528

(-0.64)

5 Greece

0.297**

(2.70)

6 Italy

0.141 (1.54)

7 Norway

0.100 (1.09)

8 Poland

-0.00724

(-0.09)

9 Spain

0.0168

(0.20)

10 United Kingdom

-0.166*

(-2.04)

𝛽𝛽

0

-0.123***

-0.388***

-0.357***

(-7.34)

(-4.09)

(-3.43)

R

2

_0.00

_0.07

_0.08

N

10200

1_{* significant at 5%, ** significant at 1%, *** significant at 0.1%} 2_{t statistics in parentheses}

3 _{Educational level (yes): Model includes dummies for 13 educational levels, level 1 as reference level}

4_{Country (yes): Model includes dummies for 10 countries, The Netherlands as reference country}

5 _{Parent*Country: Interaction variable between parent and country, The Netherlands as reference country}

6_{See appendix 1 for complete table}

(20)

Even though there does not exist a motherhood wage penalty for Dutch women it is interesting to see how the Netherlands is doing in comparison to the other involved countries. Table 5 shows the country-specific net effects of having children on female earnings. The Netherlands is omitted from the model as reference country. The individual results are therefore relative to the Netherlands.

Table 5. Net effect of parenthood for all included countries, reference country; the Netherlands

As shown in table 5 only for Greece and the United Kingdom there is a significant interaction effect.Therefore, these countries are the only counties that have a motherhood effect that differs significantly from the Netherlands. For the United Kingdom a motherhood wage penalty of -0.1124 standard deviation exists, which is in line with the findings of Joshi (1991). For Greece the effect of motherhood is 0.3505 standard deviation, which means there is no motherhood wage penalty for Greek women. Having children has actually a positive effect on the wage of Greek women. Since there is no evidence for the existence of a motherhood penalty in the Netherlands, this is also not the case for the other countries with an insignificant interaction between parent and country. According to these findings hypothesis 3 can be partly accepted. The United Kingdom is indeed in a worse situation than the Netherlands. However, Greece has a completely opposite position than expected. Looking at maternity leave length of Belgium (15 weeks), Denmark (18 weeks), France (16 weeks) and Spain (16 weeks) it is realistic that these countries show a similar effect as the Netherlands (16 weeks).

Country Parent coefficient

Netherlands (reference) Interaction effect country specific Net effect Belgium

0.0536(0.88)

0.0972(1.10)

0.1508

Czech Republic

0.0536(0.88)

-0.0668(0.79)

-0.0132

Denmark

0.0536(0.88)

-0.0990(-1.13)

-0.0454

France

0.0536(0.88)

-0.0528(-0.64)

0.0008

Greece

0.0536(0.88)

0.297(2.70)** 0.3506 Italy

0.0536(0.88)

0.141(1.54)

0.1946

Norway

0.0536(0.88)

0.100(1.09)

0.1536

Poland

0.0536(0.88)

-0.00724(-0.09)

0.0464

Spain

0.0536(0.88)

0.0168(0.20)

0.0704

United Kingdom

0.0536(0.88)

-0.166(-2.04)* -0.1124 1_{* significant at 5%, ** significant at 1%, *** significant at 0.1%}

2_{t statistics in parentheses}

(21)

According to the outcome of regression (2), there is no motherhood wage penalty, so there is no evidence for hypothesis 1 to be correct. However, it is not very strange that this result occurs. Looking at the correlation between parenthood and the control variables shown in table 6, it becomes clear that having children is correlated with the same variables that positively influence hourly wage. However, what is interesting to conclude from the correlation matrix of regression (2) is that job status is positively correlated with parent but negatively correlated with stdwage. This indicates that being a mother is correlated with working part-time and working part-part-time has a negative effect on the hourly wage. Therefore, the motherhood wage penalty could be assigned to job status and not directly to having children. This is in line with the findings of Waldfogel (2017) and Budig and England (2001) that current part-time work contributes to the motherhood wage gap

Parent Age Jobstatus Seniority Spouse stdwage

Parent 1.0000 (.) Age 0.3903*** (0.000) 1.0000 (.) Job status 0.1256*** (0.000) 0.0653 *** (0.000) 1.0000 (.) Seniority 0.1795*** (0.000) 0.4975 *** (0.000) -0.0025 (0.8005) 1.0000 (.) Spouse 0.2204*** (0.000) 0.0882*** (0.000) 0.0690*** (0.000) 0.0639*** (0.000) 1.0000 (.) Stdwage 0.0337*** (0.0007) 0.0932*** (0.000) -0.0298** (0.0026) 0.1195*** (0.000) 0.0480*** (0.000) 1.0000 (.) 1_{* significant at 5%, ** significant at 1%, *** significant at 0.1%}

2_{t statistics in parentheses}

3_{Educational levels are left out since they are categorical} Table 6. Correlation matrix for regression (2)

To give a further analysis on the exceptional outcome for Greece, table 7 is provided. As can be seen from table 7, the survey data on Greece contained a very small sample of employed women (Number of female: 378). According to World Bank statistics less than 50 percent of all Greek women participated in the labour force1_{. In addition, the use of formal childcare services is only} 20 percent2_{, which means that childcare has to be provided by the parents or informal childcare} services. Even though Greece has a relatively large number of weeks assigned to maternity leave (43 weeks) mothers from Greece are in the most favourable situation. It might be the case that the women, who actually work, have different characteristics than the average women of the included countries. In combination with the fact that parent is correlated with other control

1_{Except for Italy, Greece has the lowest female labour force participation, appendix 2}

2_{Except for Poland and the Czech Republic, Greece has the lowest use of formal childcare services,} appendix 3

(22)

variables that affect hourly wage, this might be an explanation why there is a positive effect of having children for Greek women.

Table 7 does not show any exceptional data on part-time work for the United Kingdom, which is in contrast to the findings of Joshi (1991) who claimed that the United Kingdom had the highest number of women who worked part-time at that time. According to the survey data, the United Kingdom does not have the highest percentage of women who work part-time. However, it is remarkable that despite a female labour force participation of 56%3_{, which is relatively high} in comparison to the countries used for this research, the use of formal childcare services is only 27%4_{. In line with the findings of Joshi (1991) the limited use of childcare services might be a} possible explanation why the United Kingdom does have a motherhood wage penalty relative to the Netherlands.

Table 7. Statistics on part-time work for women of all countries based on PIAAC survey data. As table 7 shows, the Netherlands has the highest percentage of mothers who work part-time, which does also hold for the overall percentage of part-time working Dutch women. There does not exists a motherhood wage penalty for Dutch women, however current part-time work does affect hourly-wage. Since many Dutch mothers seem to choose to work part-time, this can be seen as a self-selection into “mother-friendly” jobs. On the other hand it might be the case that because of this high number of women who work part-time, the wage difference between Dutch women is not that big. Only 16.35 percent of Dutch mothers in this sample works full-time. However, the effect of part-time employment on Dutch women is insignificant.

3_{Appendix 2: Female labour force participation rate}

4_{Except for the Czech Republic and Poland, the United Kingdom has the most remarkable relation}

between the use of childcare services and female labour force participation, appendix 4

Nmbr.

female Mothers Part-time Total %part-time Mothers part-time %Mothers part-time

Belgium 919 744 411 44,72% 373 50,13% Czech rep. 774 598 94 12,14% 77 12,88% Denmark 1282 1112 346 26,99% 307 27,61% France 1271 1044 387 30,45% 346 33,14% Greece 378 281 79 20,90% 60 21,35% Italy 586 394 174 29,69% 131 33,25% Netherlands 965 740 710 73,58% 619 83,65% Norway 1070 925 300 28,04% 275 29,73% Poland 956 535 107 11,19% 51 9,53% Spain 849 604 245 28,86% 189 31,29% United knd. 1150 877 548 47,65% 445 50,74% Total 10200 7854 3896 38,20% 2873 36,58%

(23)

6. Conclusion and Discussion

This research paper investigates whether having children has a negative effect on the wage of Dutch women and how the relative situation of the Netherlands is, compared to several other European countries. In order to find the effect of children on women's earnings, two different models are tested. Model (2) is used to find if there exists a motherhood wage penalty for the average European women, based on 11 European countries. The third model is used to find the effects of having children specific for the countries included. With these models, OLS-regressions are done to find out the effect of having children on a woman’s hourly wage.

Like the results imply, overall there appears to be no motherhood wage penalty. Hypothesis 1, “There does exist a motherhood wage penalty for the included European countries", is therefore rejected. This contradicts to the findings of Waldfogel (1997) and Budig and England (2001). Though, an interesting correlation has been found between part-time work, motherhood and hourly wage. Part-time work is positively correlated with motherhood, at the same time, part-time work is negatively correlated with the hourly wage. Just as Waldfogel (1997) and Bundig and England (2001) explained in their research, part-time work contributes to the existence of a wage gap between childless women and mothers. However, the coefficient job status provided with the OLS-regression (2) is not significant.

The second hypothesis, “There does exist a motherhood wage penalty for Dutch women”, also needs to be rejected. The result shows that having children has no significant effect on Dutch women’s wages. It is possible that there does not exist a motherhood wage penalty for Dutch women, because the 16 weeks of maternity leave that are allowed in the Netherlands prevent women from staying out of the labour market for too long, just as Gupta and Smith (2007) use as an argument. In addition, the Netherlands has a large share of mothers who work part-time, according to Waldfogel (1997) and Bundig and England (2001) this might be a possible explanation for a wage gap between childless women and mothers. It can be the case that this does not hold since the number of Dutch women who work part-time is very large. However, the coefficient for job status provided with OLS-regression (3) is not significant.

Hypothesis 3, ‘The Netherlands will do better than Greece and the United Kingdom” can partially be confirmed. The results show that the Netherlands has a higher coefficient for motherhood than the United Kingdom, in line with findings of Joshi (1991) there does exist a motherhood wage penalty for women from the United Kingdom. However, Greece is doing way better than the Netherlands even though Greece is expected to have a high motherhood wage penalty. Though, Greece has a very small sample of women who actually work, it might be the case that the women who do work have different characteristics that positively affect earnings.

(24)

With these results, there can be concluded that there exists no such penalty as a motherhood wage penalty for Dutch women. Therefore, the research question ‘What is the effect of having children on the earnings of Dutch women and how is the situation of the Netherlands in comparison to other European countries?’ can be answered. Several researchers found that women have to deal with a motherhood wage penalty; while others mention that the existence of this wage gap can be assigned to different labour market choices, like the choice to work part-time or other choices that lower work experience. However, in the Netherlands there is no motherhood wage penalty. Children have no significant effect on the earnings of Dutch women. In addition, mothers from the Netherlands are in a better situation than mothers from the United Kingdom. Because of the characteristics of the data set that is used for this research, it might be the case that the result of some countries is not accurate or significant. For some countries only a small sample of women is available. Besides, it is data from a survey, which is very sensitive to human error or intentional wrong answers.

(25)

7. Limitations

There have to be acknowledged, that this research paper only provides descriptive results. The results that are found cannot determine the causal effect of having children on a women’s income. This paper only provides descriptive information on the relative motherhood wage penalty of the Netherlands. In addition, the use of survey data exposes this research to some limitations. A survey can easily provide insights on many variables but there has to be taken into account that the collection of this data is subject to human error. Besides, it is possible that the confidentiality is low. The fact that people know they are participating in a survey, might cause them to twist their answers, think about job status or income. It might be the case that people lie about their background. Because of the option to give no answer, some questions remain unanswered and therefore lots of observations can no longer be used.

However, there are options to determine the causal effect of having children. According to Kleven, Landais, and Sogaard (2017), randomised fertility would be the ideal experiment to study the effect of children. Since this is practically impossible, researchers have tried to approach this in a different way. For example, Rosenzwieg and Wopin (1980) and Bronars and Grogger (1994) did this by using twin births. Angrist and Evans (1998) on the other hand, used parental preferences to have a mixed-sex sibling combination as an instrumental variable to find the labour-supply consequences of having children. However, with this approach they could only find an effect of the third child. So, the limitation of using these instruments is that it can still not provide information on the causal effect of children in general, especially not on the effect of the first-born child. It does provide causal effects of having a second or third child. To find out what causal impact children have on the labour market paths of mothers, Kleven, Landais and Sogaard (2017) use an event study approach. This event study is based on changes around the birth of the first child. They found that women and men follow a parallel path of earnings, hours worked, participation rates, and wage rates until the first child arrives. After this event their paths diverge sharply and do not fully converge ever again, a motherhood penalty remains. By using an event study approach Kleven, Landais and Sogaard (2017) managed to plot the effect of the arrival of a first child. This can be approved as a causal effect.

In 2017 Lundborg, Plug, and Rasmussen found a way to somehow create randomisation of fertility. They examine the effect of having children on women who do not have children yet. To find the effect of a first-born child they used in vitro fertilisation (IVF) as an instrumental variable. As a natural experiment, they use women who had their first child after a successful IVF treatment at first IVF attempt. As comparison group they use women who failed to have a child after the first IVF treatment. Since the career paths of successfully and unsuccessfully treated women are practically the same and the success of IFV treatments does not depend on

(26)

labour market success, Lundborg, Plug, and Rasmussen believe that IFV treatment can create an exogenous shock to the chance to have children.

In order to actually be able to find a causal link between fertility and labour market outcome, IVF-driven fertility has to capture an exogenous shock. They found that most women, who undergo IVF, undergo multiple treatments to achieve success. Since IVF treatments are very costly financially, as well as psychological, women who do undergo multiple treatments are probably women with more resources or a stronger demand for children. Using women who undergo more IVF treatments can make the effect IVF endogenous. Therefore, Lundborg, Plug, and Rasmussen take only women with a first attempt into account. To find an overall effect on children they only consider women that are childless when they start their first treatment. In this way IVF treatment success can create exogenous variation in the probability to have children, therefore IVF can be used as an instrumental variable to determine a causal effect of having children different labour market outcomes.

Lundborg, Plug, and Rasmussen (2017) had because of access to a very original dataset the possibility to conduct this experiment. They used administrative data on IVF treated women in Denmark. Beside the IVF register, they used longitudinal information on demographic variables and labour and market variables. These two datasets were merged so that they had the unique opportunity to run their experiment.

With the use of their instrumental variable, they found that having children has a negative effect on labour market outcomes. Besides, this negative effect is especially driven by the change in hourly earnings. In addition, they found that women move to lower-paid jobs, closer to home, after they have children.

These two examples show that it is possible to find a causal effect of having children. By finding a way to use the first-born child as an exogenous shock, a causal effect can be found. However, this is not easy to find, Lundborg, Plug, and Rasmussen (2017) therefore contributed to this field of research in a completely new way.

(27)

Reference list

(Journals in blue are on the Tinbergen list)

Angrist, Joshua D., and William N. Evans. 1998. “Children and Their Parents’ Labor Supply: Evidence from Exogenous Variation in Family Size.”

American Economic Review 88 (3): 450–77.

Becker, G. (1985). Human Capital, Effort, and the Sexual Division of Labor. Journal of Labor Economics, 3(1, Part 2), S33-S58.

Becker, G (1991). A Treatise on the Family. Cambridge, MA: Harvard University Press.

Blank, R. (1990). "Are Part-Time Jobs Bad Jobs?" Pp. 123-55 in A Future of Lousy Jobs? The Changing Structure of U.S. Wages, edited by G. Burtless. Washington, DC: Brookings Institute.

Bronars, Stephen G., and Jeff Grogger. 1994. “The Economic Consequences of Unwed Motherhood: Using Twin Births as a Natural Experiment.” American Economic

Review 84 (5): 1141–56.

Budig, M., & England, P. (2001). The Wage Penalty for Motherhood. American

Sociological Review, 66(2), 204-225.

CBS (2014). Gender pay gap: Fact or fiction? Retrieved from https://www.cbs.nl/en- gb/news/2014/47/gender-pay-gap-fact-or-fiction-

CBS (2016). Krijgen mannen en vrouwen gelijk loon voor gelijk werk? Retrieved from

https://www.cbs.nl/nl-nl/nieuws/2016/47/krijgen-mannen-en-vrouwen-gelijk-loon- voor-gelijk-werk-

Davies, H. & Joshi, H. (1995). "Social and Family Security in the Redress of Unequal Opportunities.".The Economics of Equal Opportunities. 313-34.

Eurostat (2016). Labour market and Labour force survey (LFS) statistics. Retrieved from: http://ec.europa.eu/eurostat/statisticsexplained/index.php?title=Labour_marke t_and_Labour_force_survey_(LFS)_statistics#Effect_of_having_children

Fransen, E., Plantenga, J. & Vlasblom, J.D. (2012). Why do women still earn less than men? Decomposing the Dutch gender pay gap 1996-2006. Applied Economics,

44(33),4343-4354.

Goldin, Claudia. (1990). Understanding the Gender Gap: An Economic History of American Women. Oxford, England: Oxford University Press.

Goldin, C. (2014). A Grand Gender Convergence: Its Last Chapter †. American Economic

Review, 104(4), 1091-1119.

Gupta, N., & Smith, N. (2002). Children and Career Interruptions: The Family Gap in

(28)

Hill, M. (1979). The Wage Effects of Marital Status and Children. The Journal of Human

Resources, 14(4), 579.

Jacobsen, Joyce P., & Levin, Laurence M. (1995). Effects of Intermittent Labor Force Attachment on Women's Earnings. Monthly Labor Review, 118(9), 14-19.

Joshi H. (1991). Sex and motherhood as handicaps in the labour market, in Groves D., M. Maclean (Eds.) Women’s Issues in Social Policy. London: Routledge.

Kleven, H., Landais, H. & Søgaard, J.E. (feb 2017). Children and Gender Inequality: Evidence from Denmark. The National Bureau of Economic Research (NBER). Klerman, J., & Leibowitz, A. (1999). Job continuity among new mothers. Demography,

36(2), 145-155.

Korenman, S., & Neumark, D. (1992). Marriage, Motherhood, and Wages. The Journal of

Human Resources, 27(2), 233.

Lundborg, P., & Plug, E., Rasmussen, A.W. (2017). Can women have children and a career? IV evidence from IVF treatments. American Economic Review, 107(6),

1611-1637.

McLanahan S.S., Kelly E.L. (2006) The Feminization of Poverty. In: Handbook of the Sociology of Gender. Handbooks of Sociology and Social Research. Springer, Boston, MA

Miller, C.C. (13 may 2017). The Gender Pay Gap Is Largely Because of Motherhood. The

New York Times. Retrieved from

https://www.nytimes.com/2017/05/13/upshot/the-gender-pay-gap-is-largely- because-of-motherhood.html

Miller, C.C. (5 feb 2018). Children Hurt Women’s Earnings, but Not Men’s (Even in Scandinavia). The New York Times

https://www.nytimes.com/2018/02/05/upshot/even-in-family-friendly- scandinavia-mothers-are-paid-less.html

Mincer, J., & Polachek, S. (1974). Family Investments in Human Capital: Earnings of Women. Journal of Political Economy, 82(2, Part 2), S76-S108.

Naur, M. and Smith, N. (1997). Cohort differences in the gender wage gap. I. Persson and

C. Jonung, Women’s Work and Wages. London: Routledge, pp. 122–44.

Neumark, David, & Korenman, Sanders. (1994). Sources of bias in women's wage equations: Results using sibling data. Journal of Human Resources, 29(2), 379.

OECD (2018). About the Survey of Adult Skills (PIAAC). Retrieved from http://www.oecd.org/skills/piaac/aboutpiaac.htm (30 june 2018)

OECD (1999). Classifying Educational Programmes. http://www.oecd.org/education/1841854.pdf Rosenfeld, R., & Kalleberg, A. (1991). Gender Inequality in the Labor Market.

(29)

Acta Sociologica, 34(3), 207-225.

Rosenzweig, Mark R., and Kenneth I. Wolpin. 1980. “Life-Cycle Labor Supply and

Fertility: Causal Inferences from Household Models.” Journal of Political Economy 88 (2): 328–48.

Rosholm, M., & Smith, Nina. (1996). The Danish gender wage gap in the 1980s a panel data study. Oxford Economic Papers, 48(2), 254-279.

Van der Meer, P. (2008). Is the gender wage gap declining in the Netherlands? Applied Economics. 40(2), 149-160.

Waldfogel, J. (1997). The Effect of Children on Women's Wages. American Sociological

Review, 62(2), 209-217.

Waldfogel, J. (1998). Understanding the "family gap" in pay for women with children. The Journal of Economic Perspectives : EP: A Journal of the American Economic

(30)

Appendix

Appendix 1: Complete table of OLS-regressions

OLS Estimates – motherhood wage penalty

Standardised

(1)

(2)

(3)

wage

Parent

0.0653***

0.0493*

0.0536

(3.41)

(2.37)

(0.88)

Age

0.00695***

0.00662***

(6.29)

(5.94)

Job status

-0.0239

-0.0246

(-1.34)

(-1.36)

Seniority

0.00906***

0.00898***

(8.34)

(8.28)

Spouse

0.0608**

0.0565**

(3.11)

(2.85)

Educational level

1

0.000

0.000 (.)

(.)

2 -0.292**

-0.315**

(-2.93)

(-3.16)

3 -0.282***

-0.300***

(-3.46)

(-3.67)

4 -0.192*

-0.208*

(-2.27)

(-2.45)

5 -0.169*

-0.182*

(-2.14)

(-2.31)

6 -0.145

-0.161*

(-1.84)

(-2.05)

7 -0.0838

-0.0987

(-0.94)

(-1.11)

8 -0.129

-0.137

(-1.26)

(-1.33)

9 -0.116

-0.127

(-0.95)

(-1.05)

10 -0.0393

-0.0564

(-0.37)

(-0.54)

(31)

11 0.0227

0.00892

(0.29)

(0.11)

12 0.158*

0.143 (1.97)

(1.78)

13 0.347***

0.333***

(4.32)

(4.13)

14 0.607***

0.583***

(5.44)

(5.23)

Country

1 Belgium

-0.136***

-0.217**

(-3.53)

(-2.71)

2 Czech republic

-0.0538

-0.00480

(-1.33)

(-0.06)

3 Denmark

-0.111**

-0.0269

(-3.04)

(-0.34)

4 France

-0.117**

-0.0754

(-3.24)

(-1.01)

5 Greece

-0.0630

-0.287**

(-1.27)

(-2.98)

6 Italy

-0.0179

-0.113

(-0.42)

(-1.46)

7 Norway

-0.331***

-0.421***

(-8.93)

(-5.03)

8 Poland

-0.136***

-0.136*

(-3.46)

(-2.04)

9 Spain

-0.172***

-0.185*

(-4.50)

(-2.54)

10 United Kingdom

-0.00701

0.116 (-0.19)

(1.60)

11 Netherlands

0.000

0.000 (.)

(.)

(32)

Interaction

Parent*Country

1 Belgium

0.0972

(1.10)

2 Czech Republic

-0.0668

(-0.74)

3 Denmark

-0.0990

(-1.13)

4 France

-0.0528

(-0.64)

5 Greece

0.297**

(2.70)

6 Italy

0.141 (1.54)

7 Norway

0.100 (1.09)

8 Poland

-0.00724

(-0.09)

9 Spain

0.0168

(0.20)

10 United Kingdom

-0.166*

(-2.04)

11 Netherlands

0.000 (.)

𝛽𝛽

0

-0.123***

-0.388***

-0.357***

(-7.34)

(-4.09)

(-3.43)

R

2

_0.00

_0.07

_0.08

N

10200

1_{* significant at 5%, ** significant at 1%, *** significant at 0.1%} 2_{t statistics in parentheses}

3_{Educational level 1 is reference level} 4_{The Netherlands is reference country}

(33)

Appendix 2: Female labour force participation rate

Source: World Bank (2012), labour force participation rate, female (% of female population ages 15+)

Appendix 3: Use of formal childcare services

Source: Eurostat (2012),

children

in formal childcare or education by age group and duration - EU-SILC survey 0% 10% 20% 30% 40% 50% 60% 70%

Labour force participation rate, female

(% of female population ages 15+)

(34)

Appendix 4: Labour force participation vs. use of formal childcare services

Relation between female labour force participation rate and use of formal childcare services, 2012 (World Bank, Eurostat)

0% 10% 20% 30% 40% 50% 60% 70% 80%

labour force participation formal childcare services

The Motherhood penalty for Dutch women : a comparison between the Netherlands and other European countries