UNIVERSITY OF AMSTERDAM
Gender wage gap in the
Netherlands
Bachelor Thesis
Niels Westenberg 28-6-2015
1
Abstract
This thesis aims to provide an answer to the question: “Is there a gender wage gap in the
Netherlands?” Data from the Centraal Bureau van de Statistiek shows that women are more and more high educated and have more years of experience in the labor market. According to the theory, this should increase their net hourly wage. To test this, an OLS regression will be done with data from the Dutch Labor Supply Panel from 2010. To give answer to this question, there were made three hypothesizes. We find that there is indeed a positive correlation between years of schooling and net hourly wage and between years of experience and net hourly wage for women. But the last
hypothesis, there is no gender wage gap, does not hold. We find that there is a gender wage gap of 8.231%, even after controlling for years of education and tenure. This means that the net hourly wage of men is 8.231% higher than that of women.
2 Statement of Originality
This document is written by Student Niels Westenberg who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is 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 of completion of the work, not for the contents.
3
1 Introduction
Every year The United Nations publishes a list with the Gender Equality Index of all the participating countries. According to this list, the differences between men and women are very small in the Netherlands for the past few years. Within the 142 participating countries, the Netherlands is ranked number 17, 15, 11, 13 and 14, respectively, from 2010 till 2014. The Netherlands had a score of 0.7608. This means that women are being treated equally for 76.08% in comparison to men. But compared with Iceland, number one country on the list with a score of 0.8276, this is not even that bad. The calculation is based on maternal mortality, the number of women in parliament, education and the employment rate of the working population (Human Development Reports, 2014). However, the research study of the United Nations did not take wages of men and women in account.
According to Groot and Groot (2004, p.2) there is a rising wage inequality in OECD countries. But they say The Netherlands was seen as an exception to this trend.
The gender wage gap in the Netherlands has not been discussed much recently. The last gender wage gap study of the Netherlands dates back from 2004. Albrecht, van Vuuren and Vroman (2004, p. 34) find that there is a gender wage gap of almost 17.6%. If the theory of Groot and Groot is correct, the difference in net hourly wage between men and women should decrease over time in the Netherlands. Therefore, this paper will investigate the gender wage gap in the Netherlands and addresses the question that: Is there a gender wage gap in the Netherlands? I adopt the Mincer wage equation to conduct the econometric analyses.
The results show that there is a gender wage gap in the Netherlands. Our regression results concluded that there is a gender wage gap of 8.231% and women still have a lower net hourly wage than men. That is smaller than there was in 2004.
The rest of the paper is structured as follow. Chapter two will include all the relevant economic theory. In the next chapter, chapter three, the hypothesizes will be listed and the model will be build that is needed for this thesis. After that, the data will be sorted. This means dropping all the irrelevant information from the dataset. In chapter four the regression will be done with the corrected dataset and framework. Chapter five will discuss the results and that will lead to a conclusion. The last chapter will discuss the shortcomings, if there are any found, for further research.
4
2 Literature review
2.1 The gender wage gap explained
In this chapter the economic theory behind the gender wage gap will be discussed. First, the two processes will be explained which can cause inequality in wages. This is needed to understand the importance of the gender wage gap. After that, general causes of the gender wage gap will be listed. This will be done by splitting up both processes and deal with them separately. Later in this chapter there will be looked at trends in the literature in the Netherlands to see how this influences the gender wage gap.
According to Gilliland (1993, p.700), there are procedural justice rules and distributive justice rules of perceived fairness. One of the distributive rules is equality. This means that all employees, without discriminating by gender, should have the same opportunity or chance of receiving the outcome (p.718). There will be no further focus on discrimination. Therefore, the gender wage gap can be explained by two processes. The first one is gender-specific factors (Albrecht et al., 2003, p.146). These are gender differences in either “qualifications or labor market treatment of similarly qualified individuals”. The second process is pure wage inequality (Blau & Kahn, 1992, p. 533). Blau and Kahn stated that wage inequality happens when there are different prices set on various labor market skills. If the difference between a high skilled worker and low skilled worker is the same in both countries, but the reward is higher in one country there will become a wage gap. If there is inequality in one of the two or both processes, there will be a gender wage gap. The bigger the inequality between both sexes, the bigger the gap.
The first process means, according to Albrecht et al., that the characteristics of men and women are not the same. For example, men are more muscular than women, which makes them more able to do physical work. Other differences in qualifications that are measurable and important for the model that will be used, are experience and education. So how can the gender wage gap be explained according these characteristics?
For many years women used to have less labor market experience and tend to have a lower education than men (Becker, 1985, p. 36). In the past it was normal for women to drop out of school and help the family. This has a negative impact on the level of education. As a result of the lower level of education the wages for women would drop. Also years of experience in the labor market had to deal with this problem. When any family had, for example, health issues the daughter was the one who was looking after the family. Women are much more caring than men (Guberman, Maheu & Maillé, 1992, p. 607). Looking after the family will make women have to quit their job, while men will
5
continue working. However, some women do not quit their job. The problem is that they cannot put all their energy in their job. Taking care of the household must be seen as a part-time job. Putting a lot of effort in taking care of the household will make them tired for their paid job and will reduce productivity (Becker, 1985, p.43).
Another effect which can cause a gender wage gap is fertility. The impact of children on the income of women is discussed multiple times. When the woman is pregnant, she can work less or cannot work at all, because she need to rest. Most of the time, the man will continue to work full-time, or will make even longer days to compensate for the loss in income of the woman. But because women know at some point that they will have children, they do not invest in themselves as much as men do (Blau & Kahn, 2000, p. 6). This will influence their educational level as well. When the women have given birth to their babies, the women do not return to the old work schedule they had before pregnancy immediately. Reason of this is because women will stay home to take care of their children. As shown figure 2.1, there is a negative relation between occupational income of the family and number of children. The more children they have, the less occupational income.
Figure 2.1. Fertility by Occupational Income of a family converted in 2000 Dollars using CPI in US.
Source: Jones & Tertilt (2008)
Another argument to explain the gender wage gap in the past, is that women are less mobile than men (Loprest, 2002, p. 527). In traditional families have to cook so they cannot work till late or make long days. Hoisl (2007, p. 620) concludes that mobility is positively related with productivity. A
6
decrease in mobility means a decrease in productivity. From this can be concluded that the average wage for women is below the average wage for men.
All these reasons have an impact on educational level and experience in the labor market. From those influences can be concluded that the gender wage gap is affected over time. This is in one line with what Becker stated, namely that “the earnings of women are adversely affected by household responsibilities even when they want to participate in the labor force as many hours as men” (1985, p.43).
The second process is pure wage inequality (Blau & Kahn, 1992, p. 533). When the difference between a high-skilled job and the low-skilled job is smaller than the difference in pay, you get paid too much for the extra skill that is needed. If this occurs, there will be a wage gap. This is the so called skill price, because you get paid according to your skill. This wage gap has increased dramatically over the past decades.
One of the factors that increased the skill price of men is the level of technology. The personal computer made its entrance to the workforce. The increase in technology boosted the demand for high-educated workers in the world (Katz, 1999, p.15). As explained earlier, on average men were better educated than women. Because of the higher demand, the price of labor increased, which means that the wages increased for men. So the change from pencil to computer work had a big impact on the wages for men (Card & di Nardo, 2002, p.26) and causes the wage inequality to rise.
Another development that has occurred is globalization. Especially in Europe, nations are cooperating with each other. Not only political rules with rules for governments, but also rules for trade. Trade barriers were partly or completely removed between the participating countries. This increased the mobility of products, including labor (Feenstra & Hanson, 1996, p.1). The labor market expanded from one nation to multiple nations across the border. This increased the competition for labor prices, the so called skill price (Groot & de Groot, 2011, p. 3). As the theory has shown earlier, on average, women were less educated and had less experience than men. That means that most of the women were low-skilled workers. Due globalisation there was a big increase in the supply of low-skilled workers across nations. This will make the wages of the low-skilled workers drop, which affect mostly the average net hourly wage of women.
To summarize this part, the gender wage gap can be explained by difference in the characteristics of men and women and by pure wage inequality. Women tend to have a lower education level and years of experience, because they spend more time with the family than men and drop out of school. This causes their average wages to drop. Also shifts in demand and supply for high-skilled workers and low-skilled workers had an impact on the wage level of men and women. There was a demand increase for high-skilled workers, because of the introduction of technology,
7
which were mostly men. The pay level of men increased. However, there was also a supply boost of low-skilled workers due globalization. This causes the average pay level of women to decrease.
2.2 Trends
This seems pretty bad for the gender wage gap. But times are changing. Since 2000 there are big changes in the characteristics of women in the Netherlands. One of them is schooling. Before 2000, as explained earlier, women were less educated than men. Women took care of the family, while men were investing in themselves to get a good job and earn money for the family. Education was not seen as right for everybody. Figure 2.2 will show the number of people who are high educated. High educated means that the person has finished high school or university in the Netherlands , as described by Centraal Bureau voor de Statistiek.
Figure 2.2. Education Dashboard. Source: Centraal Bureau voor de Statistiek
There was indeed a gap between number of men with a degree and women with a degree in the Netherlands. On average, men were higher educated than women from 1996 until 2009. Since 2010 more women have a degree in the Netherlands, compared to men in that period. This is due that there is a big increase in the beginning of 2001 till 2014 for women, and this increase is much smaller for men. This has consequences for the wage of women. The difference in education between men and women was narrowing and is even diverging in the favor of women. A higher education level will
500 600 700 800 900 1000 1100 96 97 98 99 01 01 02 03 04 05 06 07 08 09 10 11 12 13 14
Siz
e x 1
0
0
0
Year
Number of people with a degree in the
Netherlands, age 15-65
men women
8
increase the gross wages (Martins & Pereira, 2003, p. 367). This means that the wages of men and women are growing to each other. To see if this statement holds for the Netherlands, this hypothesis will be tested later.
But also the labor force participation of women increased. Not only because the education level of women increased, but also because the distribution of the household changed over time. Most of the household is still done by the woman. But cooking for the family and cleaning the house is now divided more equally between men and women (Hook, 2006, p. 654). Women have more time to work now, because they are less tired for their paid job. Figure 2.3 will give an impression of the increasing participation of women in the Netherlands.
Figure 2.3. Labour force participation rates. Source: Jaumotte (2003).
In a twenty year timespan, the participation rate in the Netherlands almost doubled. From 40 percent in 1981 to almost 80 percent in 2001. This can also be addressed to the fact that there is a labor demand change. When the technologic development started to rise, there was a demand for high educated workers. These were mostly men. But since women also become high educated, the demand for high educated women started to increase as well (Welch, 2000, p. 447). The demand shifted from physical skills to intellectual skills, due the increasing technology. So the said skill price of intellectual work increased and so did the wages for women.
Not only has changed the distribution of labor of the household, but also the composition of the household. There is a mixed theory about fertility and income. Jones, Schoonbroodt and Tertilt
9
(2008, p. 57) argued that indeed there is a negative relationship between income and child rate, as stated before. However, they are not sure what the causation is. Is it because that people with high income have low child rate or that because of low child rate they have a high income? For this research the causation does not matter. What does matter is the correlation. According the Centraal Bureau voor de Statistiek, the number of children has decreased over the years. This can be seen in Table 2.1. Table 2.1. Rate of Children. Period Children 2000 2013 Total 206619 171341 Average men 1,614 1,638 Average women 1,723 1,679
Source: Centraal Bureau voor de Statistiek (2015).
From 2000 till 2013 there is a decrease of 206619 - 171341 = 35278 children. This is a decrease of almost seventeen percent. If the rate of income between 2000 and 2013 is added to this table, the correlation can be measured between number of children and total income. The correlation is provided in Table 2.2.
Table 2.2.
Correlation between Income and Rate of children.
income kids income Pearson Correlation 1 -.937**
Sig. (2-tailed) .000
kids Pearson Correlation -.937** 1 Sig. (2-tailed) .000
**. Correlation is significant at the 0.01 level (2-tailed). Source: Centraal Bureau voor de Statistiek (2015).
There is indeed a big negative correlation between rate of children and total income. According to this table and the previous theory, there can be concluded that there was a rise in income.
The recent trends show that the wages of women are increasing and therefore narrowing the gender wage gap. Could it be that there is no gender wage gap anymore? To check the theory, the following hypothesizes will be tested with a model:
10
𝐻1: 𝐻𝑖𝑔ℎ𝑒𝑟 𝑒𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛 𝑖𝑠 𝑐𝑜𝑟𝑟𝑒𝑙𝑎𝑡𝑒𝑑 𝑤𝑖𝑡ℎ ℎ𝑖𝑔ℎ𝑒𝑟 𝑖𝑛𝑐𝑜𝑚𝑒 𝑓𝑜𝑟 𝑤𝑜𝑚𝑒𝑛. 𝐻2: 𝑀𝑜𝑟𝑒 𝑒𝑥𝑝𝑒𝑟𝑖𝑒𝑛𝑐𝑒 𝑖𝑠 𝑐𝑜𝑟𝑟𝑒𝑙𝑎𝑡𝑒𝑑 𝑤𝑖𝑡ℎ ℎ𝑖𝑔ℎ𝑒𝑟 𝑖𝑛𝑐𝑜𝑚𝑒 𝑓𝑜𝑟 𝑤𝑜𝑚𝑒𝑛. 𝐻3: 𝑇ℎ𝑒𝑟𝑒 𝑖𝑠 𝑛𝑜 𝑔𝑒𝑛𝑑𝑒𝑟 𝑤𝑎𝑔𝑒 𝑔𝑎𝑝.
3. Mincer Model
The model that will be used to research the hypothesizes is the Mincer wage equation. The Mincer wage equation is the most used economic framework to determine wages (Björklund & Kjellström, 2002, p.195) and is named after Jacob Mincer. The Mincer wage equation takes the logarithm of wage as a function of the sum of years of schooling, experience in the labor market and quadratic experience in the labor market. This is known as a log-linear model and a one-unit change in a parameter goes along with a 100% times corresponding parameter change in the logarithm of wage (Stock & Watson, 2012, p. 312).
Schooling is measured in years from the beginning of the school career, till the person gets a degree or drops out. Card and Krueger (1990, p. 41) found in their paper that schooling has a significant impact on wages. They conclude that there is a positive relationship between level of schooling and the wage level. But maybe even more important, they also found that this relationship is linear (p. 48).
Mincer also found that there is a positive relationship between experience in the labor market and level of wage. However, according to the human capital theory, the return of experience on wages is declining over time (Leminieux, 2006, p. 135). Therefore the relationship is non-linear rather than linear. This can be explained with the example of an old person. In the beginning the person will begin at the bottom of the wage distribution. When the person is getting older,
experience will increase and so does productivity. But the additional experience you get after a long time does not increase your productivity that much anymore. The marginal productivity decreases, but is still positive. Because it still does increase your wage, Mincer used two parameters that include experience. One is the linear relationship which is increasing, the second parameter is quadratic experience, which should have a negative effect to correct for the declining return in experience over time.
The only thing that must be added to complete the Mincer wage equation is a dummy variable. A dummy variable is added to capture the difference in earnings between men and women. It will have the value of 1 if the person is a man and 0 if the person is a woman.
At last, control variables must be added to the equation to capture other characteristics to perform the best possible regression. In table 3.1, in attachment one, there is a brief description of all the variables that are used and needed in this regression.
11
If we combine the Mincer wage equation with the dummy variable and control variables, we get this formula:
log 𝛾 = 𝛽0+ 𝛽1∗ 𝑆𝑐ℎ𝑜𝑜𝑙𝑖𝑛𝑔 + 𝛽2∗ 𝐸𝑥𝑝𝑒𝑟𝑖𝑒𝑛𝑐𝑒 + 𝛽3∗ (𝐸𝑥𝑝𝑒𝑟𝑖𝑒𝑛𝑐𝑒)2+ 𝛽4∗ 𝑀𝑎𝑙𝑒 + 𝛼1∗
𝐴𝑔𝑒 + 𝛼2∗ (𝐴𝑔𝑒)2+ 𝐶𝑜𝑛𝑡𝑟𝑜𝑙 𝑉𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠
4. Data
The data that is used for this research comes from the Labor Supply Panel in the Netherlands. It is founded by the OSA (Organization for Strategic Labor market Studies). This is a long-term survey among the Dutch population and their work situation. Most of the people that are interviewed are part of the Dutch labor force. That means that these are people from 15 till 65 years old, who can work, will work and are allowed to work. This data comes from 2010 and contains 4872 records, which is the sample size. This data set has all the information that is needed to continue with this research. After filtering out those who are self-employed, retired, or students, the data set has 2714 records left. Not all the information is perfect for use. Binary variables have to created and some transformation of variables is needed as well. For example, the name of the degree must be translated into years of schooling. Table 4.1 provides a list with the name of education in Dutch and the average years of schooling a person needed to finish to get their degree, calculated from primary school.
Table 4.1.
Years of total Schooling per education.
Education Total years of Schooling
VMBO 13.5 HAVO 14 VWO 14.5 MBO 20 HBO 22 WO 24
Source: Centraal Bureau voor de Statistiek (2015).
Last thing that is needed to do a proper regression is to convert the net wage in one month, to net hourly wage and take the logarithm of this. In the end, this data set includes 1311 women and 1403 men.
12
5. Results
In order to give answers, the first thing that will be done is looking at the average wage for men and women to get a global perspective. In the following table, Table 5.1, you can see that men earn €13.73 on average, while women earn €12.35 on average. That is a difference of €1.38. Using a t-test we found a t-value of 8.1, which is bigger than the critical value of 1.96. This means that this
difference is significantly different from zero. There are also bigger outliers for men. In this survey the highest net hourly wage of men was €85.94 and for women it was €31.25. But also the lowest net hourly wage was earned by women, €1.46. In the case of men, the lowest net hourly wage was €4.54. If the gender wage gap is calculated by the means, there will be found a gender wage gap of 11.2%.
Table 5.1.
Net hourly wages in Euro’s by gender.
Wages
Gender Min Max Average Std. Error
Men €4.54 €85.94 €13.73 5.17
Women €1.46 €31.25 €12.35 3.44
This seems to be a big gender wage gap and that is the reason a regression is needed to test whether the estimated parameters are significantly different from zero. In table 5.2 in the attachments, the regression output is described.
What do these numbers mean? Since there is used a logarithm-linear model, the value of the coefficient needs to be multiplied by 100% to see what a one unit change of that variable will do with the output in per cent. For example, one additional year of schooling will result in a 0.041*100% change in the net hourly wage for women. You have to keep in mind to check if this value is significantly different from zero. There is used a 95% confidence interval, so the t-value of that coefficient needs to be larger than |1.96|. We see a t-value of 20.46, so this value is significantly different from zero.
5.1 Higher education is correlated with higher income for women.
The first hypothesis that will be tested will check if there is a positive correlation between years of schooling and net hourly wage for women. In chapter two we saw the trends of high educated women in the Netherlands. The number of high educated women is increasing over time. So if there
13
is a positive correlation between these two, the gender wage gap is narrowing over time in the Netherlands. In table 5.2 we find the coefficient of schooling with a value of 0.04147. As described above, the t-value of schooling is bigger than |1.96|. So this means that 0.04147 is significantly different from zero. Since there is a logarithm-linear model used, a one unit change in years of schooling will result in a 0.04147*100%=4.147% increase in net hourly wage for women.
5.2 More experience is correlated with higher income for women.
This hypothesis states that the more experience you have in the labor market, the higher your net hourly wage is as a woman. In this regression there are two variables where years of experience is involved, Experience and Experience2. As you can see, the value of the coefficient of Experience is 0.00728. This is indeed significantly different from zero, because the corresponding t-value is bigger than |1.96|. So that means that one additional year of experience in the labor market, results in a 0.00728*100%=0.728% increase in the net hourly wage of women. However, if we look at the coefficient of Experience2 we find a negative value. This value is -0.00003. But by checking the corresponding the t-value, we find that |-0.53| is not bigger than |1.96|. So this is not significantly different from zero and does not influence the net hourly wage of women. Because one additional year of experience increases the net hourly wage of women, we can say that there is indeed a positive correlation between experience and income.
5.3 There is no gender wage gap.
By looking at the theory we might expect no gender wage gap at all. But if we calculated the gender wage gap by looking at the means, we come to the conclusion that there is a gender wage gap of almost 11%. Therefore, there is made a dummy variable which catches the difference in net hourly wage between men and women. This dummy is added in the OLS regression and is called “Male”. By looking at table 5.2 we find the coefficient of Male with a value of 0.08231. To check if this value is significantly different from zero, we compare the t-value of |2.72| with the critical t-value of |1.96|. The t-value of Male is bigger than the critical value, so this coefficient is significantly different from zero. That means that if you are male, you net hourly wage will increase with 0.08231*100%=8.231%. So the difference in net hourly wage between men and women, called the gender wage gap, is 8.231%.
14
6. Conclusion
This thesis had its main goal to provide an answer to the main question: “Is there a gender wage gap in the Netherlands?” According to Groot and Groot the gender wage gap in the OECD countries was going to increase. However, they said that the Netherlands were an exception to this trend. Albrecht, van Vuuren and Vroman find that there is a gender wage gap of almost 17.6% in 2004. From the theory we find that the number of women who are high educated is increasing. Also women got more experience in the labor market than before 2000, because the division of the household is split more equally between man and woman. That should not have increase the net hourly wage of women by definition. That is why there are made three hypothesizes that will help answering the main question and check if the theory of Groot and Groot is correct. The first hypothesis is that higher education is correlated with higher income for women. From the OLS regression output we can conclude that that is indeed the case. Women with more years of schooling, and therefore higher educated, will have a 4.147% net hourly wage than women with one year less of schooling. The second hypothesis is that more experience is correlated with higher income for women. This hypothesis is correct as well. From table 5.2 we can conclude that women with one year additional experience will receive a 0.728% higher income than women with one year less experience. The last hypothesis states that there is no gender wage gap in the Netherlands at all. That means the net hourly wage of men and women are the same. If we look at the dummy “Male” we can see that this hypothesis has to be rejected. The t-value of “Male” is bigger than the critical t-value. That means there is a gender wage gap of 8.231% in the Netherlands. But according the Groot and Groot the gender wage gap should be smaller over time. If we check that, we see that that statement is correct. It was 17.6% in 2004 and with using the data of 2010 we find a gender wage gap of 8.231%. Since the first two hypothesizes are correct, this gender wage gap might decrease even more over time.
7. Suggestions for further research
There are some suggestions to improve the regression after all. If we summarize the net hourly wages of men and women, we find that there are big outliers. This is shown in table 5.1. These outliers can be measurement errors. In further research you might check the database more carefully. Some impossible outliers, like negative net hourly wage, were already removed.
The next suggestion is to check for multicollinearity between the parameters. All the
“Working Sector” dummies affect the net hourly wage of men in a negative way. The ones that have a positive effect on net hourly wage for men are not significant. But some of them are
male-15
dominated sectors, like the constructional and transport sector. Zellner (1972, p. 157) says that in male-dominated occupations, the relative wages of men are significantly higher than that of women. But in this regression it is the other way around. If you work in a male-dominated sector as a men, it decreases your net hourly wage.
Since years of schooling has a positive correlation with net hourly wage and experience has a positive correlation with net hourly wage as well, there can be done further research on these topics. When people go to school, in general, they do not work, so schooling has a negative correlation with experience. Therefore, one can research if there is an optimum amount between years of schooling and years of experience that yields the highest return to net hourly wage.
16
Preliminary reference list of journals from the Tinbergen Journal List
Albrecht, J., Björklund, A., & Vroman, S. (2003). Is There a Glass Ceiling in Sweden? Journal of Labor
Economics, 21(1), 145-177.
Arulampalam, W., Booth, A. L., & Bryan, M. L. (2007). Is There a Glass Ceiling over Europe? Exploring the Gender Pay Gap across the Wage. Industrial and Labor Relations Review, 60(2), 169-173. Becker, G. S. (1985). Human Capital, Effort, and the Sexual Division of Labor. Journal of Labor
Economics, 3(1), 34-44.
Björklund, A.,& Kjellström, C. (2002). Estimating the return to investments in education: how useful is the standard Mincer equation? Economics of Education Review, 21, 195–210.
doi: 10.1016/S0272-7757(01)00003-6
Blau, F. D., & Kahn, L. M. (1992). The Gender Earnings Gap: Learning from International Comparisons.
American Economic Review, 82(2), 533-538.
Blau, F. D., & Kahn, L. M. (2000). Gender differences in pay. (Working Paper No. 7732). Retrieved January 11, 2015, from
http://www.nber.org/papers/w7732
Card, D., & di Nardo, J.E. (2002). Skill biased technological change and rising wage inequality: some problems and puzzles. Journal of Labor Economics, 20(4), 1-49. doi: 10.3386/w8769 Card, D., & Krueger, A. (1990). Does school quality matter? Returns to education and the
characteristics of public schools in the United States. (Working Paper No. 3358). Retrieved January 11, 2015, from http://www.nber.org/papers/w7732
Feenstra, R. C., & Hanson, G. H. (1996). Globalization, outsourcing and wage inequality. (Working Paper No. 5424). Retrieved June 11, 2015, from http://www.nber.org/papers/w5424 Loprest, P. J. (1992). Gender differences in wage growth and job mobility. The American Economic
Review, 82(2), 527-528.
Martins, P. S., & Pereira, P. T. (2004). Does education reduce wage inequality? Quantile regression evidence from 16 countries. Labour Economics,11(3), 360-371.
Welch, F. (2000). Growth in women's relative wages and in inequality among men: One phenomenon or two? American Economic Review, 90(2), 447-449.
17
Zellner, Harriet. (1972). Discrimination Against Women, Occupational Segregation, and the Relative Wage. American Economic Review, 62(2), 157-160.
Other relevant references
Charles, M. (2011). A world of difference: international trends in women's economic status. Annual
Review of Sociology, 37, 355-359.
Economic Dashboard. Retrieved January 3, 2015, from http://www.cbs.nl/nl-NL/menu/themas/onderwijs/cijfers/kerncijfers/onderwijsdashboard.htm
Gilliland, S. W. (1993). The perceived fairness of selection systems: An organizational justice perspective. Academy of management review, 18(4), 694-734.
doi:10.5465/AMR.1993.9402210155
Groot, S. P., & Groot, H. L. F. (2011). Wage inequality in the Netherlands: Evidence, trends and
explanations. CPB Netherlands Bureau for Economic Policy Analysis.
Guberman, N., Maheu, P., & Maille, C. (1992). Women as Family Caregivers: Why do they care? The
Gerontologist , 32(5), 607.
Hoisl, K. (2007). Tracing mobile inventors—the causality between inventor mobility and inventor productivity. Research Policy, 36(5), 619-625.
Hook, J. L. (2006). Care in context: Men's unpaid work in 20 countries, 1965–2003. American
sociological review, 71(4), 650-655.
Human Development Reports. (2014, January 31). Retrieved October 24, 2014, from http://hdr.undp.org/en/content/gender-inequality-index-gii
Jaumotte, F. (2003). Female Labour Force Participation: Past Trends and Main Determinants in OECD Countries.
Jones, L. E., Schoonbroodt, A., & Tertilt, M. (2008). Fertility theories: can they explain the negative
fertility-income relationship? (No. w14266). National Bureau of Economic Research.
Jones, L. E., & Tertilt, M. (2008). An economic history of fertility in the United States: 1826–1960 (pp. 165-230). Emerald Group Publishing Limited.
18
Katz, L. F. (1999). Technological Change, Computerization, and the Wage Structure. UnpublishedManuscript, Department of Economics Harvard University.
Lemieux, T. (2006). The “Mincer equation” thirty years after schooling, experience, and earnings (pp. 127-145). Springer US.
Petersen, T., & Morgan, L. A. (1995). Separate and Unequal: Occupation-Establishment Sex Segregation and the Gender Wage Gap. American Journal of Sociology, 101(2), 329-365. Rate of Children (2015). Retrieved January 3, 2015, from
http://statline.cbs.nl/StatWeb/publication/?VW=T&DM=SLNL&PA=37422ned&D1=0,4-5,7,9,11,13,17,26,35,40-41&D2=0,10,20,30,40,(l-4)-l&HD=090218-0953&HDR=G1&STB=T Stock, J. H., & Watson, M. M. (2012). Introduction to Econometrics (pp. 309-312). Essex:
Pearson.
Years of Schooling (2015). Retrieved June 11, 2015, from
http://www.cbs.nl/nl-NL/menu/themas/onderwijs/publicaties/artikelen/archief/2009/2009-2916-wm.htm
19
Attachments
Attachment 1
Table 3.1Summary of variables used in regression
Name Description Measured
zschooling Years of schooling Years
zexperience zexperience2 Years of working experience Experience*Experience Years Years zage zage2 Years of age Age*Age Years Years
zmale Dummy for Male 1 for Male
zdumpart_man Interaction term for Male and Part-time worker
1 for Male and working Part-time zdumpub_man Interaction term for Male
and Public Sector
1 for Male and working in Public sector
zdumtrans_man Interaction term for Male and Transport Sector
1 for Male and working in Transport sector zdumtrade_man Interaction term for Male
and Trading Sector
1 for Male and working in Trading sector
zdumindus_man Interaction term for Male and Industrial Sector
1 for Male and working in Industrial sector zdumhealth_man Interaction term for Male
and Healthcare Sector
1 for Male and working in Healthcare sector zdumeduc_man Interaction term for Male
and Educational Sector
1 for Male and working in
Educational sector zdumconstruct_man Interaction term for Male
and Construction Sector
1 for Male and working in
Construction sector zdumbusiness_man Interaction term for Male
and Non-public Business Sector
1 for Male and working in Non-public Business sector
zdumother_man Interaction term for Male and the Sector that does not correspond with the others
1 for Male and working in a different sector than listed above
20
Attachment 2
Table 5.2.
OLS Regression output with standard error between brackets.
Coefficient p-value Schooling 0.04147 (0.00203) 0.000** Experience 0.00728 (0.00293) 0.013* Experience2 -0.00003 (0.00006) 0.594 Age 0.032241 (0.00526) 0.000** Age2 -0.00034 (0.00006) 0.000** Male 0.08231 (0.03024) 0.007** Part-time*Male 0.12917 (0.01962) 0.000** Transport*Male -0.08402 (0.03884) 0.031* Trade*Male -0.14033 (0.03884) 0.000** Public*Male 0.00775 (0.03515) 0.825 Industrial*Male -0.05646 (0.03328) 0.090* Healthcare*Male -0.10646 (0.03927) 0.007** Educational*Male -0.07594 (0.03742) 0.043** Constructional*Male -0.09239 (0.04046) 0.022** Business*Male 0.03363 (0.03295) 0.308 Constant 1.01404 (0.09671) 0.000** Note: *p< .05. **p< .01.
