The effects of gender inequality on economic growth

(1)

The effects of gender inequality on

economic growth

Victoria F. Estrada Ploegmakers (10841059) Supervisor: Professor Cenkhan Sahin

Faculty of Economics Master’s Thesis -July 2018- Abstract:

This paper studies the impact of gender inequality in the dimensions of healthcare, education, labour force and democracy on economic growth for OECD countries for the period between 1990 and 2015. The results were obtained using the GMM dynamic panel data (DPD) method due to high endogenity between the variables. The results show that the gender wage gap in the labour force and the female labour force does not have an effect on economic growth. The secondary education of women and the life expectancy of women have a positive and significant effect on economic growth. Finally, the results also show that the number of seats in national parliaments held by women has a positive and significant impact on economic growth. However, when adding time and country dummy variables, none of the variables are significant at a 10% significance level, although the coefficients are similar. Therefore, additional investigation is needed. The results obtained in this study are important to further develop policies that aim at achieving an egalitarian society.

(2)

1.I

NTRODUCTION

In the past decades, there has been a decrease in matters of gender inequality in the dimensions of education, labour force, life expectancy and representation in democracy, as women are more empowered than before. All around the globe, women have more rights and more economic opportunities than several decades ago. However, regardless of the improvements made throughout the years, women do not benefit from the same treatment as their male counterparts in matters of healthcare, education, market labour force and politics (OECD, 2017). Many researchers have investigated the gender inequality subject. Gender equality is not only a human right, but it seems to have an impact on economic growth1_{. A large body of evidence has found that gender inequality}

in the society can have a negative impact on economic growth (Klasen, 2000; Dollar & Gatti, 1999). A more egalitarian society seems to enhance economic growth. More specifically, a higher life expectancy of women, a higher attainment of secondary education by women, an increase of the female labour force, decrease in the gender wage gap and a greater share of seats held by women in national parliaments seem to have a positive impact on economic growth. Considering the previous factors, the current literature has shown the following. Better female healthcare has a positive impact on economic growth (Bloom, Kuhn & Prettner, 2016); an egalitarian education for both boys and girls enhances economic growth (Benavot, 1989); gender inequality on the labour force affects negatively economic growth (Wolszczak-Derlacz, 2013) and finally, greater political representation by female enhances economic growth (Jayasuriya & Burke, 2013). However, there is also enough evidence supporting the argument that gender inequality in the labour force affects positively economic growth (Seguiono, 2000). Therefore, researchers have been unable to exactly determine the effects that gender inequality can have on economic growth. In order to shed some light on the subject, this paper will try to demonstrate that there is a negative relationship between gender inequality and economic growth.

This paper will try to determine the impact of gender inequality on economic growth by doing a regression that will study the impact of the secondary education of women, shares of the seats in parliament held by women, the labour force participation

1_{The United Nations Charter, signed in 1945, establishes equal rights for men and women. Moroever, it was aggreed}

(5)

of females, the gender wage gap and the life expectancy of females on economic growth in OECD countries for the period 1990-2015. This paper contributed to the current research in a newly manner as it includes different dimensions of gender inequality (i.e. healthcare, education, labour force and democracy) in order to study its impact on economic growth. Moreover, the data used is quite recent, and therefore, the concluding results will be updated to the current economic environment. The importance of this study lies in the fact that this research can further contribute to achieving a more egalitarian society in the OECD countries. By quantifying the impact of gender inequality on economic growth, new policies can be developed in order to diminish and abolish gender inequality in society. Therefore, it is crucial to keep investigating the effects of gender inequality on economic growth.

The generalized method of moments (GMM) will be used given that the method controls for the endogenous problem that may arise between the variables2. The rationale of the endogeneity problem is the following: If there is gender inequality in education, it will be very hard for women to overcome the gender gap in the labour market and in the political sphere. Better education for women can have two positive effects on the society. First, it can improve the healthcare of women leading to a decrease of gender inequality in healthcare. In return, the productivity of females in the labour force will be substantially increased, the fertility rate will be decreased, and future generations will be significantly better nourished (Feinstein, Sabates, Anderson, Sorhaindo & Hammond, 2006). Second, it will allow that a higher ratio of females enter the political arena, which in turn, will promote policies that diminish gender inequality in the different faces; healthcare, education, labour market and political representation (Kabeer, 2005). Conversely, if there is gender inequality in the labour market, girls can abstain of getting a higher education given the unfairness of the labour market towards women (Gertler & Alderman, 1989). This in turn, will lead to gender inequality in healthcare and political representation. Consequently, the reverse causality between variables works both ways. In order to study the impact of gender inequality on economic growth, this paper will run a GMM regression using the dynamic panel data method (DPD). The econometric specification will include a set of variables retrieved from Paul Romer (1986) economic growth model combined with variables that quantify

2_{In order to get the results, this paper will use the two-step linear GMM estimator., which gives robust results.}

(6)

gender inequality on the above-mentioned dimensions. The data used in this study will be taken from the World Bank online database (www.worldbank.org) and from the OECD online database (http://stats.oecd.org ). This paper has selected to study the OECD countries given the close relationship of the author to the countries. Moreover, considering the availability of the data, the period between 1990 and 2015 will be examined.

This study shows that a higher life expectancy for women enhances economic growth. Moreover, the results conclude that secondary education of women in OECD countries has a positive and significant impact on economic growth. Conversely, the study finds that the secondary education of men and the life expectancy of men do not have a significant impact on economic growth at a 10% significance level. Furthermore, the number of seats in parliament held by women has a significant impact on economic growth. However, according to the results obtained, the presence of the gender wage gap in the labour force and the female labour force does not have a significant effect on economic growth3. Nevertheless, when accounting for time and country fixed effects, the results show that none of the variables of interest are significant at a 10% significance level. However, similar coefficients for the regressors are observed when accounting for country and time fixed effects.

This thesis is structured as follows. First, a literature review of the most recent work in the field will be presented. It will include a thorough explanation of the work done in the following four sub-topics: economic growth and the gender inequality in healthcare, education, labour force and political representation. Then, section 3 will present the empirical analysis, which will include an analysis of the data. Afterwards, section 4 will present the results of the analysis. Section 5 and section 6 will test the robustness of the results and the policies that can be implemented given the results of the analysis, respectively. Finally, the last section will conclude.

3_{Klasen and Lamanna (2009) state the complexity of finding suitable instruments for the female labor force, and}

(7)

2. L

ITERATURE

R

EVIEW

This section will explain the different aspects in which gender inequality is present in societies across countries. First, the literature of the relation between economic growth and gender inequality in healthcare will be reviewed. Second, a thorough explanation of the relation between economic growth and gender inequality in education is provided. The final subsections explain the relationship between economic growth and the gender inequality in both the labour market and in the political sphere. The above-mentioned dimensions where the gender gap is present (i.e democracy, healthcare, education and labour force) have been chosen to study their impact on economic growth given that there is evidence showing their presence in OECD countries (OECD, 2014; Murtin et al., 2017; OECD, 2017; OECD, 2005)

2.1ECONOMIC GROWTH AND THE GENDER INEQUALITY IN HEALTHCARE.

A large body of literature found that a very important aspect of gender inequality lies in the healthcare that women receive throughout their lifespan. There is enough evidence showing that better female healthcare has a positive impact on economic growth through the following four channels (Bloom, Kuhn & Prettner, 2015). First, better health for women will lead to an increase in productivity given the increase in the labour supply which in turn, will positively affect economic growth. Second, through a decrease in the probability of morbidity and in the mortality rate, women get a higher return on their education investment. That is, better health affects positively the labour market participation both at the intensive and at the extensive margin4 (Jayachandran & Lleras-Muney, 2009; Albanesi & Olivetti, 2014). Third, according to Field, Robles and Torero (2009) a mother’s health is directly related to the children’s health by providing nourishment to the foetus, and later by breastfeeding and taking care of them. Consequently, better female health has a direct impact on future generations and therefore, on the country’s economic growth (Bhalotra and Rawlings, 2011). Finally, better health for women decreases the fertility rate due to the higher availability of contraceptives and hence, it increases the productivity and the labour supply, which ultimately will have a positive effect on economic growth (Bailey, 2006)

4_{In labour economics, the intensive margin is defined as how much does a person work (i.e how many hours)}

(8)

It is clear from the above-mentioned reasons that female healthcare has an impact on the country’s economic growth. Therefore, this paper will take females’ life expectancy - as a proxy for female healthcare - into account in the regression in order to test if the impact on economic growth is significant.

Given the above-mentioned arguments, this paper poses the following hypothesis:

H1: With a higher life expectancy for women, higher economic growth is expected.

2.2ECONOMIC GROWTH AND GENDER INEQUALITY IN EDUCATION

Many attempts have been made with the purpose of determining what generates growth. It has been suggested that one of the key determinants of economic growth is the investment in education (Barro, 1996). In fact, Besley & Persson (2011) found that countries will have a faster growth when they have more-educated leaders. In many countries, however, girls do not have the same education as boys given that the education gap is still present.

Some preliminary work was carried out by Becker in 1975, where he explained some of the reasons that lead to an under-investment in girls’ education. He stated that parents, as rational human beings, would be better off investing in their son’s education rather than in their daughter’s education, as boys are likely to work more years than girls. That is, parents look for the highest internal return of earnings to education when deciding how much to invest in their children’s education. Gertler & Alderman (1989) found three main features to explain such preference regarding education. First, similar to Becker’s argument, parents may take into account the return on their investment in education, and therefore they prefer investing in the boy’s education. This argument is valid when considering the wage gap in the labour market. Second, parents are securing an income when they grow old. Boys are more prone than girls to give them money because the latter one will become part of a new economic household unit and therefore, won’t be able to contribute. Third, parents may prefer boys to girls in matters of education. In some countries, this view is still consistent. Evidence shows that in many countries, the female college enrolment is much lower than male college enrolment

(9)

(Jacobs, 1996). If the evidence presented is validated, then a significant part of the population could be more educated that it actually is, which leads to an underinvestment in human capital. This can have disastrous consequences for economic growth.

Becker’s arguments have some flaws and are subject to criticism. For example, Schultz (1993) pointed out that Becker’s argument does not take into consideration the productivity of non-working women on social matters. Moreover, Benavot (1989) indicated that economic growth is larger when investing in girls’ education rather than in boys. Becker’s view fails to explain the fact that the number of highly educated women has been increasing for the past years regardless of the economical adversities that they may later find. Mickelson (1989) has suggested three explanations of this growing trend. Female groups that empower women, the possibility of meeting a high-class educated husband and false future expectations regarding the labour market are among the reasons of this upward tendency regarding female education. From this, it can be concluded that parents no longer engage in favouritism when investing in their children’s education (Hauser & Kuo, 1995; Behrman et al., 1986). Furthermore, secondary education is mandatory in many countries around the world. If this law is enforced, there won’t be any discrimination between boys and girls and both sexes will get the same secondary education5 (Jacobs, 1996). Consequently, more women can obtain a college degree, which has a positive impact on economic growth.

A large amount of research has been done in order to determine the non-monetary benefits of educating women. Investing in the education of women can lead to lower fertility rates and better health for their own children, which can have a positive impact for future economic growth (see section 2.1). Furthermore, there is a causal relationship between gender inequality in education and economic growth. The rationale is the following: higher education for women leads to an increase in national income which in turn, results in higher gender equality in education, healthcare, labour market and political representation.

On the basis of the preceding discussion, this paper poses the following hypothesis:

5_{Even
though
girls
and
boys
get
the
same
amount
of
years
of
secondary
education,
the
latter
one
is
more
likely}

(10)

H2: With a higher rate of educated women, higher economic growth is expected.

2.3ECONOMIC GROWTH AND PROFESSIONAL INEQUALITY

For the past decades, the importance of gender inequality on the labour market has emerged leading towards a greater effort to protect discriminated groups (i.e women) in the labour market (Arrow, 1972, p. 3; Becker, 1990, p. 43; Stiglitz, 1973, p. 287). For example, the United Nations Organization, OECD, International Labour Organization and the European Union, among others, have focused on narrowing the gender gap in the labour market.

There is a rapidly growing literature on how professional inequality may affect economic growth. For the sake of understanding, this paper will define gender inequality on the labour market as disadvantages (i.e. employment, promotion and remuneration) that women face in the working environment.

The current literature, however, has failed to reach a consensus on whether gender inequality in the labour market – as a form of wage gap- has a negative impact on economic growth. On the one hand, reducing the wage gap will have a positive impact on future generations, and therefore future economic growth. The rationale that sustains this statement is the following: a decrease in the wage gap will lead to an increase in the female labour force participation rate. Their income will be mostly spent on their children’s education and health, which is found to be an investment into future generations, as a more productive and competent labour force will be built for the forthcoming years, affecting positively future economic growth (Wolszczak-Derlacz, 2013). Another valid argument against the gender wage gap states that income differences between the two sexes will have an impact on the fertility rate (World Bank, 2011). If women earn lower wages than their male counterparts, the opportunity cost of having children decreases leading to an increase of the number of children per women. Consequently, the population growth will increase significantly resulting in lower economic growth.

(11)

On the other hand, some prominent economists found evidence that economic growth can be stimulated by the gender wage gap. For example, Seguino (2000) found that export countries could have a competitive advantage due to the gender wage gap. In order to export goods, a labour intensive workforce is needed. If a certain group of labour were cheaper (e.g. women), then the country would decrease their export prices leading to an increase in price competitiveness and expanding its export boundaries. In turn, the country could invest in more advanced technology - thanks to the benefits from the export expansion- and therefore, stimulate growth.

Gender inequality in the labour market persists - to varying degrees- in every OECD country even though in most OECD countries women have achieved the same level of education as their male counterparts. In fact, England (1993) reported that for the past century, working females have attained a higher median of years of schooling than men. Further evidence supporting England’s (1993) claim is found in the paper written by Gornick & Jacobs (1996), where he stated that males and females earn different wages despite having the same level of education. He also found evidence that disregarding the educational level; the wage gap does not vary. In other words, there is a wage gap in every level of education (Bernhardt et al, 1995).

In some developing countries, nevertheless, the wage gap is triggered by the education gap leading to a reverse causal relationship between both variables (Adcock, 2013). For example, if girls are not able to attain an intermediate education they will not be able to get a job in the formal sector, as employers prefer hiring a highly educated workforce. On the contrary, if the girls’ progenitors are aware of the gender wage gap in the workforce, they may abstain of investing in the girls’ education, leading to an underinvestment in their education. Therefore, the gender inequality in education and the gender gap in the labour force are largely linked.

Given the above-mentioned arguments, this paper will test if the gender inequality on the labour force- taking the wage gap as a proxy- has a negative impact on economic growth. The following hypothesis will be tested:

(12)

2.4ECONOMIC GROWTH AND POLITICAL REPRESENTATION

All around the globe, women do not participate as actively in politics as men do. In fact, women hold less than the 20% of the national parliaments’ seats (World Bank, 2012). Evidence presented by Jayasuriya & Burke (2013) shows that a greater female representation in politics generates faster economic growth. That is, empowering females in politics has a positive effect on economic growth.

Rodrik (2000) propounds the view that in order to attain sustainable economic growth it is fundamental the existence of democracy and hence, democratic institutions. According to Rodrik’s view, the markets function efficiently when people trust that their transactions will be protected by law. This protection, nevertheless, can only be obtained by assuring the government’s stability and their capability to enforce the law. With such guarantee, market players will have the certainty that their property is secured and markets will work efficiently. From the above-mentioned argument, it is clear democracy promotes economic stability, yet in most countries there is no such thing as a perfect democracy. As previously mentioned, women are hardly present at the political sphere, which leads to an under-representation of the female population. The lack of diversity in parliaments challenges the principles of every modern democracy (Caul, 1999). In order to empower women in the political sphere, social structures should change (Adcock, 2013).

An increasing number of studies have found evidence that women are less likely to engage in corrupt behaviour than men (World Bank, 2001; Swamy, Azfar, Knack & Lee, 2001; Branisa et al., 2013). It has also been demonstrated that is significantly harder to bribe a female in a position of power than a male (Swamy, Knack, Lee & Azfar, 2001). On the basis of the preceding evidence, some researchers have suggested that it would be economically beneficial to increase the number of female representatives in national parliaments as it would lead to a decrease in corruption levels and hence, to a fairer democracy in the short and medium term (Barro & Lee, 1996). Eventually, economic growth would be positively affected. Along similar lines, there is evidence showing that as more females have an active role in politics, a more egalitarian society is achieved. Women focus their attention on different policies than men. The former cares more about healthcare programs and other basic needs programs (Aulette,

(13)

Wittner & Blankley, 2009). Moreover, women are more interested than men in decreasing the gender inequality gap in sectors such as healthcare, education, and the labour market.

More recent evidence highlights that the policies taken by the government have a significant effect on economic growth (Barro, 2013). Therefore, empowering women in the political arena can have a positive effect on economic growth.

Figure 1 shows the percentage of national seats of parliament held by females in the OECD countries. As it can be seen from the following graph, the share of seats held by women has increased since 1998, leading to a higher share of female representation in national democracy. The red line shows the percentage of females holding a national seat in parliament for every OECD country in 1998, whereas the blue line shows the percentage of national seats in parliament held by females in every OECD countries in 2015. As it can be concluded from the graph, the percentage of females holding a seat in national parliament in every OECD has increased throughout the years.

Note: Percentage of national seats of parliament held by females in the OECD countries. The red line shows the percentage of females holding a national seat in parliament for every OECD country in 1998. The blue line shows the percentage of national seats in parliament held by females in every OECD countries in 2015

Source: Author’s own calculations. Data are retrieved from the World Bank.

Given the existing literature on the topic, this paper poses the following hypothesis:

(14)

H4: With a greater share of parliament seats held by women, greater economic growth is expected.

3. D

ATA AND METHODOLOGY ANALYSIS

In order to test the above-mentioned hypotheses and study the impact of gender inequality on economic growth, this paper will consider the theory developed by Paul Romer (1986) and other prominent economists to study the cross - section long-run economic growth6. Therefore, the variables needed to study the long-run economic growth will be the following: population growth (PGROWTH), investment (INV), technology (ICT) and trade (TRADE). The regression will be combined, however, with the variables that have been selected to study the impact of gender inequality on economic growth. These are the following: female labour force (LFF), gender wage gap (GWG), secondary education of women (SEW), the number of seats held by women in national parliaments (DEMOCRACY), the life expectancy of women (LEW), the life expectancy of men (LEM) and the secondary education of men (SEM). The life expectancy of males and the secondary education of men have been included to avoid omitted variable bias. More specifically, both variables have been included in the regression to study the impact of female life expectancy and secondary education of female on economic growth. In addition, the regression will also incorporate year dummies and country dummies, as it is essential to capture the time fixed effects and country fixed effects. Hence, the econometric specification will be on the following form:

where GROWTH_GDP is used as an indicator of economic growth constant 2010 dollar prices; SEW is the secondary school enrolment for females as an indicator of education; LEW is the life expectancy of women at birth as an indicator of female

6_{His
model
states
that
economic
growth
is
endogenized
in
the
model,
i.e.
long
run
economic
growth
comes}

from within the model. Romer (1986) stated that in order to achieve long-‐run economic growth, investing in human capital is essential. This model came as a novelty in the field given that in the preceding years economic growth was considered to be exogenous (Solow, 1956).

(15)

healthcare; GWG stands for gender wage gap in the labour market as an indicator of the gender gap in the labour market; LFF stands for female labour force participation;

DEMOCRACY stands for democracy and is computed using the female share of seats

held in national parliament as an indicator of female political presence; PGROWTH is the percentage of population growth; INV is the investment as a percentage of GDP;

ICT stands for information and communication technology and is computed by the

number of telephone lines for every 100 habitants; TRADE is the exports and imports as a percentage of GDP used as an indicator of to what extent is the country open; SEM is the secondary school enrolment for males as an indicator of education ; LEM is the life expectancy of men at birth as an indicator of male healthcare; stands for time

dummy variables; stands for the country dummy variable; stands for the error

term and, finally, the i and t represent the country and time (yearly basis), respectively.

3.1. DATA DESCRIPTION

This paper has selected the OECD countries to carry out its study. The aforementioned economic organisation has been chosen given the close relationship of the author to the countries. Furthermore, this paper will examine the effects of gender inequality on economic growth for the period between 1990 and 2015. The data used in this paper is retrieved from the World Bank online database (www.worldbank.org) and from the OECD online database (http://stats.oecd.org ). Both databases benefit from accurate and reliable data.

Table 1 shows the descriptive statistics of the data. Table 1: Descriptive statistics

Variable N Mean Std. Dev. Min. Max.

GWG 402 16. 2269 7.8535 0.4 41.7 LFF 844 52. 3495 10.0209 23.3 79.3 DEMOCRACY 638 23.2225 10.6831 2.4 47.3 LEW 870 80.8134 2.8550 69.5 87.1 SEW 541 90.2606 7.7648 45. 08 100 LEM 884 74.8107 2.8914 67.6 81.4 SEM 516 88.6825 7.3776 53.95 100 INV 768 22.7653 3.8131 9.832 41.37 PGROWTH 881 .5965 .7705 - 2. 25 6. 01 ICT 884 44.4601 14.9698 9.79 95.95 TRADE 867 82.9866 50.5493 16.01 410.17 dGROWTH_GDP 832 0.0208 0.1219 -2.93 1.03

Note: This table shows the statistics for OECD countries from 1990 to 2015. Source:OECD database, 1990-2015 and World Bank database,1990-2015.

(16)

Table 1 shows that the annual GDP growth for OECD countries is 2.08% and that the 90.26% of females have attained a secondary education. The high number of women with a secondary education is not surprising given that in all the OECD countries secondary education is mandatory. In this case, boys and girls have the same chance of getting a secondary education regardless of their gender. In fact, on average an 82% of the population in OECD countries has attained a secondary education (OECD, 2014). This means that, on average women have greater chances of attaining secondary education than men in almost every OECD country. Table 1 shows that on average, 88.5% of males haves attained secondary education. Evidence suggests that in only two OECD countries, Germany and Switzerland, boys have a higher rate of secondary education attainment than girls. The secondary education rate for women has been included in the regression instead of the tertiary education attainment, given that in order to achieve a tertiary education, secondary education is needed. Moreover, the secondary education of women has also been included in the regression of the current literature (see Salatin & Shaaeri, 2015 ; Cabeza-García, Del Brio, & Oscanoa-Victorio, 2018; Klasen, 2002)

Only a 52.35% of the females, however, are part of the labour force. From this result it can be concluded that a high share of females do not participate actively in the labour force7. One explanation to this trend could be due to the fact that a large number of women have to take care of their families (Stoller, 1992). That is, the low share of women in the labour force supports the social role theory8. Furthermore, on average there is a gender wage gap of 16.23%, which illustrates that on average females earn a 16.23% less than their male counterparts. Moreover, women only hold a 23.22% of the seats in national parliaments, which clearly indicate a misrepresentation of females in democracy. On average, life expectancy of women exceeds 80 years, which is above of the average life expectancy of men of (i.e. 77.9 years in 2015). Table 1 shows that on average, the life expectancy of males is 75 years (i.e. 5 years lower than the life expectancy of women). Evidence shows that women have a higher life expectancy than men given that, for the last 25 years, life expectancy gap has decreased significantly

7_{Women are more prone to work in non-economic activities, such as care-work.}

8_{The social role theory states that the society believes that women are in charge of tasks in which caregiving is}

(17)

(OECD, 2011). More specifically, between 1970 and 1980 life expectancy of men and women in OECD countries had sharply increased. However, evidence shows that the increased longevity was higher for men than for women (OECD, 2017).

Figure 2 and Figure 3 represent the relation between the life expectancy of females (LEW) and the labour force of females (LFF) and the relation between the secondary education of women (SEW) and the life expectancy of women (LEW), respectively. As Figure 2 shows, the higher the life expectancy of women, the higher the female force participation will be. Moreover, Figure 2 shows that the predicted labour force participation of female increases throughout the years leading to a decrease of the female labour participation when the retirement age approaches. This result is expected.

Figure 2: Relation between LEW and LFF

Fig 2. - Scatterplot of the relationship between the life expectancy of women (LEW) and the labour force participation of females (LFF). Sources: Life expectancy of women: OECD database, 1990-2015; Labour force participation of females: OECD database, 1990-2015. The blue dots represent the labour force participation on females. The red line represents the fitted values.

Figure 3 shows the relationship between the secondary education of women and the life expectancy of women. As it can be shown in the graph, higher secondary education of women leads to a higher life expectancy of women. This shows the correlation between both variables, which is aligned with the current literature on the subject. Better education for females lead to better healthcare for females.

20 40 60 80 70 75 80 85 90 LEW LFF Fitted values

(18)

Fig 3. - Scatterplot of the relationship between the life expectancy of women (LEW) and the secondary education of women (SEW). Sources: Life expectancy of women: OECD database, 1990-2015; Secondary education of women: OECD database, 1990-2015. Blue dots represent the life expectancy of women. The red line represents the fitted values.

Figure 4 shows a histogram of the gender wage gap (GWG). As it can be shown in the graph, the data is right skewed. Therefore, the mean of the gender wage gap data is greater that the median. Moreover, Figure 4 shows that a 16.92% of the OECD countries have a gender wage gap superior to 15%. In other words, in 16.92% of the countries of the OECD, women earn approximately a 15% less than men. In an extreme case, a 2.24% of the OECD countries have a gender wage gap of 40%. That is, in 2.24% of the OECD countries, women earn on average a 40% less than men.

Figure 4: Histogram of the gender wage gap

Fig 4. – Histogram of the gender wage gap (GWG). Sources: Gender wage gap: OECD database, 1990-2015.

1.741 3.98 9.95 14.18 11.19 16.92 15.17 12.19 4.478 2.488 1.493 1.741 .4975 1.7412.239 0 5 10 15 20 Pe rce n t 0 10 20 30 40 GWG Figure 3: Relation between SEW and LEW

70 75 80 85 90 40 60 80 100 SEW

(19)

The current literature has specified the endogeneity between the variables. Female healthcare (i.e. life expectancy of females) is related to the education females receive, the labour force - consequently, the gender wage gap - and the parliament seats held by women. In order to test for endogenity, this paper will run the Durbin-Wu-Hausman Test9. This statistical test detects whether endogenity is present in the regression (Chmelarova, 2007). If that were the case, the OLS estimation will give biased results, resulting on a failed estimation. The null hypothesis of the test states that there is no correlation between the variables and the error term. After running the Durbin-Wu-Hausman Test, at a 5% significance level the null hypothesis is rejected, resulting on the existence of endogenous variables. The life expectancy of women is correlated with the gender wage gap, the secondary education of women, the labour force of females and the seats held by females in national parliaments.

The generalized method of moments (GMM) in a dynamic panel data (DPD) model will be used in order to estimate the effects of gender inequality on economic

growth on OECD countries for the period 1990-201510. The GMM was first developed

by Hansen in 1982. A large body of researchers have applied the method in different context such as: panel data, cross-sectional regressions and time series. The GMM method appears to be an appropriate estimator for data containing a large amount of individuals and for a limited time period. Moreover, GMM is appropriate when the variables are endogenous, the individuals are heteroskedatic and autocorrelated and the panel data has fixed effects (Roodman, 2006). The benefits of the method lie on the fact that it controls for endogeneity between the variables. Moreover, the GMM gives a more robust outcome than other methods. The GMM method outperforms other tests (i.e. IV estimator) if the errors are heteroskedastic. On the contrary, if the errors were homoskedastic the IV estimation would be more efficient.

9_{Durbin (1954), Wu (1973,1974) and Hausman (1978) developed a method to test for endogeneity between a}

regressor and the error term. In the Durbin-Wu-Hausman Test, the null hypothesis is specified as H0: cov (x,e)=0 and

the alternative hypothesis is specified as H1: cov (x,e) ≠0 . If the null hypothesis is not rejected, then it is the case

that the least squares estimator and the instrumental variables estimator are consistent. On the contrary, if the null hypothesis is rejected, then it is the case that the least squares estimator is not consistent, whereas the instrumental variables estimator is consistent.

10_{In order to get the results, this paper will use the two-step linear GMM estimator., which gives robust results.}

(20)

As previously mentioned, this paper will use the dynamic panel estimators by Arellano & Bond (1991), Arellano & Bover (1995) and Blundell & Bond (1998) given the fact that there is correlation between some variables and the error term. In order to solve the endogeneity problem and get unbiased results, instruments are needed. The Blundell-Bond method allows for lagged values of the regressors to be used as instruments when the model is in first difference. This paper will use the investment as a percentage of GDP (INV), the percentage of population growth (PGROWTH), the information and communication technology (ICT) and the exports and imports as a percentage of GDP (TRADE). The most common method to test the validity of the instruments is running the Sargan-Hansen test11. The null hypothesis of the Sargan Hansen test state that the instruments used in the regression are valid and therefore, they are uncorrelated with the error term. The instruments used in this paper pass the Sargan-Hansen test.

The GMM estimator relies on the ‘mean stationarity’ assumption. More specifically, stationarity implies that the mean, the variance and the autocorrelation of the data do not change over time. In order to test for stationarity, this paper will run a Fisher-type test using the augmented Dickey-Fuller test. This test is the most appropriate given that the data set is unbalanced. The null hypothesis states that all panel data contain unit roots, whereas the alternative hypothesis states that at least one panel data is stationary. After running the Fisher - test, it can be concluded that not all the variables are stationary. Therefore, the first difference of every variable will be used to correct for stationary. The results of the Fisher-test for first differenced variables show that they are stationary. The econometric specification will be on the following form:

11_{Hansen’s J test checks if there are over-identification of intruments . Under the null hypothesis, the instruments are}

valid.This implies that the instruments satisfy the following two conditions: instrument relevance, corr (Z,X) ≠0,

and instrument exogenity, corr (Z,u)=0

(21)

Multicollinearity among the variables can give biased results when running the regression. One ore more regressors can suffer from multicollinearity if they are correlated with each other. In the given case that the variables suffer from perfect collinearity, the model won’t be able to be estimated. In order to avoid estimation bias, the correlation between variables should not exceed 0.8 (Farrar & Glauber, 1967; Mason & Perreault, 1991). Table 2 shows the correlations between the variables.

Table 2: Correlations

dGWG dLFF dDEMOCRACY dLEW dSEW dLEM dSEM dINV dPGROWTH dTRADE dICT dGROWTH_GDP

dGWG 1. 0000 -0.0265 0.1680 0.0333 -0.0050 0.0155 0.0180 -0.1729 -0.0692 -0.0720 -0.1145 -0.0416 dLFF -0.0265 1. 0000 0.0306 0.0497 0.1492 0.0421 -0.0530 0.2117 0.1534 0.0134 0.1566 0.2392 dDEMOCRACY 0.1680 0.0306 1. 0000 -0.0088 0.0384 0.0371 -0.1964 -0.1353 -0.2930 -0.0575 -0.0396 -0.0501 dLEW 0.0333 0.0497 -0.0088 1. 0000 0.0800 0.0813 0.0068 -0.1592 -0.1560 0.1309 0.0931 - 0.0014 dSEW -0.0050 0.1492 0.0384 0.0800 1. 0000 -0.0822 0.1653 -0.0549 0.0646 -0.0603 0.0969 0.0764 dLEM 0.0155 0.0421 0.0371 0.0813 -0.0822 1. 0000 0.1112 0.0538 0.0527 -0.0020 -0.0353 -0.1386 dSEM 0.0180 -0.0530 -0.1964 0.0068 0.1653 0.1112 1. 0000 0.0140 0.0050 -0.0989 0.1089 -0.0006 dINV -0.1729 0.2117 -0.1353 -0.1592 -0.0549 0.0538 0.0140 1. 0000 0.5659 0.2637 -0.0752 0.4804 dPGROWTH -0.0692 0.1534 -0.2930 -0.1560 0.0646 0.0527 0.0050 0.5659 1. 0000 0.0018 0.1039 0.4502 dTRADE -0.0720 0.0134 -0.0575 0.1309 -0.0603 -0.0020 -0.0989 0.2637 0.0018 1. 0000 -0.0607 0.2048 dICT -0.1145 0.1566 -0.0396 0.0931 0.0969 -0.0353 0.1089 -0.0752 0.1039 -0.0607 1. 0000 -0.0115 dGROWTH_GDP -0.0416 0.2392 -0.0501 - 0.0014 0.0764 -0.1386 -0.0006 0.4804 0.4502 0.2048 -0.0115 1. 0000

Note: This table shows the correlation between the different variables of the regression. Source: OECD database, 1990-2015 and World Bank database, 1990-2015.

In order to control for year fixed effects, time dummy variables will be added. With the help of time dummy variables, the variables of the regressors can capture the effects of time. In other words, thanks to the time dummy variables, the variables can capture the heterogeneity between time. The function of dummy variables is to capture the unobserved effects that vary over time but are constant throughout the panel. Furthermore, dummy variables help increase the strength of the analysis by decreasing the possibility of omitted variable bias in the regression (Stock & Watson, 2015). Following this line of reasoning, country dummy variables will be also added in the regression in order to control for country effects. Country dummy variables are needed to interpret the results compared to the reference country. Consequently, equation (1) and (1.1) will contain time and country dummy variables.

(22)

4. R

ESULTS

This section will review the results obtained from the estimation of equation (1.1) in the OECD countries for the period between 1990-2015. The results in Table 3 show the following:

(23)

Table 3:GMM results. Impact of gender inequality on economic growth.

VARIABLES (1) (2) (3) (4)

dGROWTH_GDP dGROWTH_GDP dGROWTH_GDP dGROWTH_GDP

dGWG -0.00126 -0.000965 -0.000779 -0.00100 (0.000765) (0.000860) (0.000862) (0.000954) dLFF 0.00561 0.00470 0.00321 0.00411 (0.00347) (0.00413) (0.00407) (0.00485) dLEW 0.0206** 0.0213** 0.0103 0.00589 (0.00863) (0.0104) (0.0117) (0.0146) dSEW 0.00329* 0.00325 0.00101 0.000744 (0.00187) (0.00206) (0.00237) (0.00255) dSEM 0.00000 0.00000 0.00000 0.00000 (0.000194) (0.000214) (0.000235) (0.00243) dLEM 0.000408 0.00361 0.00316 0.00358 (0.00954) (0.0104) (0.0114) (0.0124) dDEMOCRACY 0.00225* 0.00234* 0.00113 0.00117 (0.00118) (0.00127) (0.00136) (0.00147) dPGROWTH 0.00799 0.00655 0.0122 0.00946 (0.00818) (0.00969) (0.0103) (0.0118) dINV 0.0104*** 0.0106*** 0.00531*** 0.00523** (0.00128) (0.00160) (0.00200) (0.00222) dICT 0.00000 0.000742 (0.00175) (0.00210) dTRADE 0.00000 -0.000188 (0.000335) (0.000414) Years 1990-2015 1990-2015 1990-2015 1990-2015

Countries OECD OECD OECD OECD

Time effects? No No Yes Yes

Country effects? No No Yes Yes

Observations 94 94 94 94 P-value tests Sargan Test 0.2761 0.3571 0.3101 0.2462 Serial correlation: 1st order 0.0001 0.0000 0.0002 0.0004 2nd order 0.5342 .7689 0.3360 0.2771

Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Note: * Statistically significant at a 10% level; ** Statistically significant at a 5% level; ***Statistically significant at a 1% level. Cross-sectional study for OECD countries for the period between 1990-2015.Data from the World Bank and OECD database.

(24)

The results of equation 1.1 show the following. In column 1, gender wage gap (dGWG) is negative but does not have a significant effect on economic growth (p = 0.100). Contrary to the expected results, the labour force participation of females (dLFF) does not have a significant impact (p = 0.107) on economic growth12. Life expectancy of women (dLEW) has a positive and significant effect (p = 0.017) on economic growth at 5% significance level. More specifically, an increase of 1% in the life expectancy of women will lead to an increase in economic growth of 2.06%, ceteris paribus. Bloom, Kuhn & Prettner (2016) found similar results showing that better female healthcare has a positive impact on economic growth. The secondary education attained by women (dSEW) in OECD countries does have a positive and significant effect (p = 0.078) on economic growth at a 10 % significance level. Similarly, Benavot (1989) found that an egalitarian education for both boys and girls enhances economic growth. Conversely, the secondary education of males (dSEM) and the life expectancy of males (dLEM) do not have an impact on economic growth at a 10% significance level. Furthermore, the number of seats held by women in national parliaments (dDEMOCRACY) has a positive and significant effect on economic growth (p = 0.057). This result is in line with the current literature, which finds that empowering women in the political arena can have a positive effect on economic growth13_{. In this model, the}

effect of the population growth rate (dPGROWTH) in economic growth is not significant (p = 0.329). Finally, investment (dINV) is significant at a 1% significance level (p = 0.000)

When adding the variable dICT and dTRADE in column 2, the results obtained are similar to the results obtained in column 1. The gender wage gap (dGWG) has a negative but not significant impact on economic growth (p = 0.262). The labour force participation of females (dLFF) does not have a significant impact (p = 0.255) on economic growth (see footnote 12). Life expectancy of women (dLEW) has a positive and significant effect (p = 0.041) on economic growth at 5% significance level. The secondary education attained by women (dSEW) in OECD countries does not have a significant effect (p = 0.116) on economic growth. Similar to column 1, the secondary education of males (dSEM) and the life expectancy of males (dLEM) do not have an

12_{Klasen and Lamanna (2009) state the complexity of finding suitable instruments for the female labor force, and}

therefore, this result may be subject to bias.

13_{More recent evidence highlights that the policies taken by the government have a significant effect on economic}

(25)

impact on economic growth (p = 0.740 and p=0.729, respectively).. Furthermore, the number of seats held by women in national parliaments (dDEMOCRACY) has a positive and significant effect on economic growth (p = 0.065) at a 10% significance level. In this model, the effect of the population growth rate (dPGROWTH), trade (dTRADE) and technology (dICT) in economic growth is not significant (p = 0.499, p = 0.883 and p = 0.975, respectively). Investment (dINV) is significant at a 1% significance level (p = 0.000)

Column 3 shows the results of the regression used in column 1 with time fixed effects and country fixed effects. That is, column 3 has time and country dummy variables. Given the limited space available to show the results, the coefficient of all the years are not shown in the table. Moreover, due to collinearity between the variables, some time dummy variables have been dropped from the regression. Also, the country dummy variables were dropped from the regression. The reason is that the fixed effects estimator already accounts for cross sectional data, so the dummy variables are omitted since they cannot explain nothing. However, this doesn’t pose a problem, as the year and country dummy variables are not variables of interests. The results of column 3 and 4 show the following. None of the variables of interest (i.e dLEW, dSEW, dLFF and

dDEMOCRACY) are significant. However, the coefficients of the variables are in line

with the results showed in column 1 and 2. In other words, the coefficients of the variables of interest do not differ from the ones of column 1 and 2. For example, the gender wage gap has a negative impact on economic growth (although not significant). The female labour force impacts positively economic growth. Finally, the secondary education of women, the life expectancy of women and the number of seats held by women in national parliament, have a positive impact on economic growth.

After running the GMM regression, the following two tests were performed to check the validity of the instruments and the autocorrelation in first-differenced errors: the Sargan-Hansen test and the Arellano-Bond test. The results of the Sargan-Hansen test showed that the instruments used are valid. The results of the Arellano-Bond test showed that there is autocorrelation of order one, but the test did not reject autocorrelation of order two.

(26)

Given the above-mentioned results, the following hypotheses are not rejected:

a) With a higher life expectancy of women, higher economic growth is expected.

b) With a higher rate of educated women, higher economic growth is expected

c) With a greater share of parliament seats held by women, greater economic growth is expected.

On the contrary, given the results obtained in this study, the following hypothesis is rejected:

d) With a higher wage gap, lower economic growth is expected

Hypothesis d is rejected because the gender wage gap does not have a significant effect on economic growth

Nevertheless, when accounting for the time and country fixed effects all the hypothesis are rejected given that none of the variables of interest are significant at a 10% significance level.

In order to test the internal validity of the results, the following section will perform a robustness check on the results obtained in this section.

5. R

OBUSTNESS CHECK

Some of the results obtained in section four are aligned with the results obtained in the current literature regarding the effects of gender inequality on economic growth. For example, this paper has established that a higher life expectancy of women, secondary education of women and the number of seats held by women in national parliament, have a positive and significant effect on economic growth. This result, as previously mentioned, is aligned with the existing literature on the matter. That is, eradicating the gender gap in these dimensions has a positive effect on economic growth. However,

(27)

some of the results obtained in the previous analysis need a closer examination. For example, given the complex task of finding suitable instruments for estimating the effect of the female labour force on economic growth, the results obtained do not entirely match with the expected results. In order to test the validity of the results, a robustness check is needed.

This paper will do a robust test running regression 1.1 with a smaller sample size. In order to do so, this paper will study the impact of gender inequality on economic growth for the period between 1995 and 2010 for OECD countries. Table 4 shows the results obtained after running the regression.

(28)

Table 4: GMM results. Impact of gender inequality on economic growth.

VARIABLES (1) (2) (3) (4)

dGROWTH_GDP dGROWTH_GDP dGROWTH_GDP dGROWTH_GDP

dGWG -0.00177** -0.000796 -0.00101 -0.00142 (0.000903) (0.00157) (0.00140) (0.00414) dLFF 0.00453 0.00564 0.00617 0.00714 (0.00488) (0.00590) (0.00746) (0.0141) dLEW 0.0275* 0.0187 0.0000 -0.00180 (0.0150) (0.0191) (0.0291) (0.0451) dSEW 0.00216 0.00215 0.00247 -0.00186 (0.00349) (0.00430) (0.00606) (0.00979) dSEM 0.00000 0.00000 0.000256 0.000249 (0.000228) (0.000283) (0.000382) (0.000605) dLEM -0.00815 -0.00605 -0.00849 -0.0151 (0.0123) (0.0147) (0.0198) (0.0328) dDEMOCRACY 0.00394** 0.00290 0.00291 0.00227 (0.00160) (0.00207) (0.00227) (0.00433) dPGROWTH 0.0118 0.00887 0.0170 0.0141 (0.0151) (0.0211) (0.0240) (0.0413) dINV 0.0104*** 0.00868*** 0.00473 0.00468 (0.00172) (0.00254) (0.00398) (0.00619) dICT 0.00114 0.00101 (0.00265) (0.00606) dTRADE 0.000516 -0.000207 (0.00052) (0.00121) Years 1995-2010 1995-2010 1995-2010 1995-2010

Countries OECD OECD OECD OECD

Time effects? No No Yes Yes

Country effects? No No Yes Yes

Observations 59 59 59 59 P-value tests Sargan Test 0.3882 0.2648 0.4607 0.4355 Serial correlation: 1st order 0.0068 0.0000 0.0053 0.0162 2nd order 0.8840 0.8990 0.8524 0.8581

Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Note: * Statistically significant at a 10% level; ** Statistically significant at a 5% level; ***Statistically significant at a 1% level. Cross-sectional study for OECD countries for the period between 1995-2010.Data from the World Bank and OECD database.

(29)

The results of equation 1 show the following. In column 1, gender wage gap (dGWG) has a negative and significant effect on economic growth (p = 0.050). This means that an increase of one per cent in the gender wage gap decreases economic growth by 0.177%, ceteris paribus. The labour force participation of females (dLFF) has positive impact but not significant impact (p = 0.353) on economic growth. Life expectancy of women (dLEW) and the number of seats held by women in national parliaments (dDEMOCRACY) has a positive and significant impact on economic growth (p = 0.066 and p = 0.014, respectively). Conversely, the secondary education of males (dSEM) and the life expectancy of males (dLEM) do not have an impact on economic growth at a 10% significance level. This result is similar to the one obtained in table 3, column 1. However, secondary education of women (dSEW) has a positive, but not significant effect on economic growth. Similar coefficients are obtained in column 2, when two additional regressors have been added (i.e dICT and dTRADE). The difference, however, is that none of the variables of interest in column 2 are significant at a 10% significance level. Column 3 and column 4 show the results obtained when adding time dummy variables and country dummy variables into the regression. The results in column 3 and 4 are also similar to the results obtained in column 3 and 4 of section 4. After running the validity check, it can be concluded that the analysis of section 4 benefits from internal validity.

When accounting for the impact of gender inequality on economic growth, there are vast differences between the countries around the world. For example, in developed countries the secondary education is mandatory for boys and for girls. In many developing countries, however, there is still a huge part of the underage population - especially girls- which are unable to attend high school (Klasen, 2002). This, in the long run, will contribute to a wider gender gap in other different dimensions (e.g. labour force, healthcare and political representation). Given the high correlation between the variables, the impact of the variables (i.e. female labour force, secondary education of women, life expectancy of women, number of seats held in parliaments by women and the gender wage gap) on economic growth will be significantly different. Cuberes & Teignier (2011) reported that depending on the geographical position, the levels of gender inequality could significantly differ.

(30)

6.P

OLICY IMPLEMENTATION

The results obtained in the previous section (see section four) can be helpful in order to support the implementation of new policies that aim at reducing the gender gap in society. It is important to develop policies which main goal is to diminish gender inequality, given that the society as whole benefits from an egalitarian society (Brighouse & Wright, 2008). From an economic perspective, it is also beneficial to diminish the gender gap in the dimensions where it is present (e.g. healthcare, education, democracy and labour force).

In the previous sections, it has been thoroughly explained the economic benefits -in terms of economic growth- of diminishing and eradicating the gender gap. For example, there is enough evidence supporting that equal healthcare for men and for women enhances economic growth. Moreover, there is recent evidence showing that both men and women receive on average a better treatment on healthcare when accounting for the gender gap in healthcare (WHO, 2009). Therefore, the society benefits as a whole when preventing the bias. The results obtained in section four show that a higher life expectancy of women increases economic growth. Given the results of the analysis and considering that health policy contributes to diminishing gender inequality (Abdool et al., 2012; Lin & L’Orange, 2010), this paper proposes the following policy implication, which is similar to the one proposed by Sarah Payne (2009). First, a tougher legislation that aims at protecting the integrity of all the patients equally needs to be developed. Second, the budget that aims at achieving an egalitarian society should increase. This measure has been successfully implemented in some European countries (Payne, 2009). Finally, it is important to educate and train the participants in the healthcare system against gender bias.

Even though this analysis has not found that the gender wage gap has significant effect on economic growth, it has found that the gender wage gap has a negative coefficient. According to the International Labour Office (2016), the estimated period of time to achieve an equal remuneration for the same work done by men and women is of 70 years. A large body of evidence has determined that equal pay for men and women leads to a more equal society (ILO,2016). Similarly, it is beneficial for the society if the percentage of females in the workforce increases, even though this paper has found little

(31)

significant evidence regarding the positive impact on economic growth. Evidence suggests that a higher share of females in the workforce can affect positively the society of the future, as women tend to spend a high share of their income in their children’s education and healthcare (Wolszczak-Derlacz, 2013). Therefore, implementing new policies to achieve equal rights in the workforce has been a topic of discussions for the past years.

New policies regarding the number of seats in national parliaments held by women should also be considered. This paper has found statistical evidence about the impact of the number of women in national parliaments on economic growth for OECD countries. Evidence suggests that a higher presence of females in national parliaments reduces corruption. Furthermore, if the analysis of section four is replicated taking the data from Asian countries, the number of seats held in parliament by women has a positive impact on economic growth (Panday, 2008). Therefore, the results obtained in the analysis of section four will still be valid if other data is taken (e.g. developing countries). These results can be used to support policies aiming at increasing the quota of women in parliaments. International organizations such as the United Nations and Inter-parliamentary Union (IPU) are working on implementing the above-mentioned measure (United Nations, 2014).

7.C

ONCLUSION

This paper has studied the effects of gender inequality on economic growth. Specifically, it has focused on the effects of the gender gap in education, healthcare, labour force and democracy on economic growth. The results have shown - when not accounting for time fixed effects and country fixed effects - that higher life expectancy of women affect positively economic growth. This result implies that a higher life expectancy of females enhances growth. This conclusion is important in order to implement policies that aim at achieving gender equality in the dimension of healthcare. Moreover, this paper has found that secondary education of women does have a positive impact on economic growth, whereas the secondary education of male does not have an impact on economic growth. Furthermore, the number of seats held by women in national parliament has a positive impact on economic growth. This result is important to implement policies that increase the number of female representatives in parliaments

(32)

as a higher female representation leads to a more egalitarian society (Aulette, Wittner & Blankley, 2009). Finally, given the results of this study, gender wage has negative, but not significant impact on economic growth and female labour force has a positive, but not significant impact economic growth. However, the latter result is subject to doubt. According to Klasen & Lamanna (2009), finding suitable instruments for determining the impact of female labour force on economic growth is a complex and troublesome task. Therefore, in order to improve and obtain more accurate results a further investigation is needed. Nevertheless, the results obtained in this analysis are robust and benefit from internal validity. More specifically, when running the regression for the period between 1995 and 2010, the results are statistically similar. Therefore, the results of this paper have concluded that the number of seats held by women in national parliament and life expectancy of women have a positive impact on economic growth.

The results obtained in this analysis can be further implemented to support the development of new policies that aim at reducing gender inequality in the different dimensions: healthcare, education, labour force and democracy. A more extensive research has to be done in this area in order to develop new policies that aim at reducing gender inequality.

(33)

8. B

IBLIOGRAPHY

Adcock, K. (2013). Globalization, Women's Empowerment and Sustainable Growth: Development Theory with a Vagina. Global Societies Journal, 1(1).

Albanesi, S., & Olivetti, C. (2014). Maternal health and the baby boom. Quantitative

Economics, 5(2), 225–269.

Alderman, H., & Gertler, P. (1989). The substitutability of public and private health care for the treatment of children in Pakistan.

Arellano, M., & Bond, S. (1991). Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. The review of economic studies, 58(2), 277-297.

Arellano, M., & Bover, O. (1995). Another look at the instrumental variable estimation of error-components models. Journal of econometrics, 68(1), 29-51.

Arrow, K.J. (1972). The Theory of Discrimination, in: O. Ashenfelter, A. Rees, A. (ed.), Discrimination in Labor Markets, 1972, Princeton.

Aulette, J. R., Wittner, J & Blankely, K. (2009). Gendered Worlds. New York: Oxford

University Press. p. 139, 305.

Bailey, M. J. (2006). More power to the pill: The impact of contraceptive freedom on women's life cycle labor supply. The Quarterly Journal of Economics, 121(1), 289-320. Barro, R. J. (1996). Determinants of economic growth: A cross-country empirical study (No. w5698). National Bureau of Economic Research.Barro, R. J. (2013). Education and economic growth. Annals of Economics and Finance, 14(2), 301-328. Barro, R. J., & Lee, J. W. (1996). International measures of schooling years and schooling quality. The American Economic Review, 86(2), 218-223.

Becker, G. S. (1975). Front matter, human capital: A theoretical and empirical analysis, with special reference to education. In Human Capital: A Theoretical and Empirical

Analysis, with Special Reference to Education, Second Edition(pp. 22-0). NBER.

Becker, G. S., Murphy, K. M., & Tamura, R. (1990). Human capital, fertility, and economic growth. Journal of political economy, 98(5, 2), S12-S37.

Behrman, J. R., & Taubman, P. (1986). Birth order, schooling, and earnings. Journal of

The effects of gender inequality on economic growth

The effects of gender inequality on

economic growth

TABLE OF CONTENTS

1.I

2.

L

R

3.

D

4.

R

5.

R

6.P

7.C

8.

B