University of Groningen
Faculty of Economics and Business
Master Thesis International Economics and Business
The Development of the Female Labor Force Participation Rate in
Former Soviet Countries
Name: Amber Vos
Student number: S2830285
Email address: a.a.vos.1@student.rug.nl
Date Thesis: 18 June, 2019
Name Supervisor: K.M. Wacker, PhD
1 Abstract
This thesis empirically examines the development of the female labor force participation rate (FLPR) of the fifteen former Soviet Union (FSU) countries. These countries consider gender equality as one of the key legacies of their socialist past. Using a cross-sectional empirical analysis of 167 countries, including the 15 FSU countries, between 1991 and 2018, I find that the dummy variable representing the former Soviet Union countries remains positive and significant, even when controlling for several economic and demographic variables. Meaning that almost 30 years after the fall of the Soviet Union, there are still lasting differences between the FLPR in the former Soviet republics and the rest of the world. These results hint towards the large impact that historical contingencies and policies promoting gender equality can have on the female labor force participation rate.
Keywords
2 1. Introduction
In economic life, half the potential of skilled labor which is one of the important elements of competitiveness, is created by women (Karoui and Feki, 2017). It is well known that economic and social development depend, among many other factors, upon a rational exploitation of human resource endowment (Austen, 2005). Gender equality matters in its own right but gender equality and increased female labor force participation are also important for the economic development of nations (Sen, 1998). Increased participation of women in the labor force has a positive impact on social and economic outcomes. It leads to productivity gains, which have become vital in a globalized and competitive world (World Development Report, 2012). Beyond economic benefits, women’s participation in the labor force can be seen as a signal of declining discrimination and increasing empowerment of women (Klasen and Pieters, 2012; Mammen and Paxson, 2002). Empowering women as both political and social actors can change policy decisions and make institutions more inclusive and representative of the whole population (Duflo, 2012; World Development Report, 2012). Understanding what influences the female labor participation rate is necessary in order to understand the implications for women’s wellbeing and for future growth of the labor force and the economy. The incorporation of women in the workplace can improve the level for overall human capital in a society, which is a prerequisite for further economic development (Amate-Fortes, Guarnido-Rueda and Molina-Morales, 2015).
Under state socialism, women fared relatively well in the labor market: female-male differentials were similar to those in the West, and female labor force participation rates in the Soviet Union were among the highest in the world (Brainerd, 2000). Table 1 shows that before the fall of the Soviet Union, in 1990, the female labor force participation rates (FLPR) in the Soviet republics were noticeably higher than the FLPR in the rest of the world. The difference between the FLPR in the Soviet states and the FLPR of the rest of the world in 1990, is as large as 5.1 percentage points.
Table 1: Average Female Labor Force Participation Rates
Group of countries Average FLPR in 1990 Average FLPR in 2018
World 51.4 % 51.5%
World excluding the
former Soviet countries 48.1 % 51.3%
Former Soviet countries 53.2 % 53.7%
Source: International Labor Organization, 2019.
3 decrease. This thesis analyzes the development of the female labor participation rate in the 15 former Soviet Countries. This includes countries from different regions: Central Asia (Kazakhstan, Kyrgyzstan, Tajikistan, Turkmenistan, and Uzbekistan), the Caucasus (Armenia, Azerbaijan, and Georgia), Eastern Europe (Belarus, Moldova, Russia, and Ukraine) and the Baltics (Estonia, Lithuania , and Latvia). The aim of this thesis is to examine the development of the FLPR in these countries and to answer the following research question: did the female labor participation rate in the former Soviet countries become more ‘normal’? Additionally, can a combination of economic and demographic variables explain the variation between the FLPR in the former Soviet republics and the rest of the world?
Recent theoretical and empirical studies suggests that women’s labor market participation unambiguously promotes growth and excluding women from economic life produces inefficiencies and limits the economic performance of countries (Luci, 2009; Klasen, 2002; Klasen, 2018). Klasen and Lamanna (2009) argue that this is because gender-specific discrimination limits the ‘talent pool’ and allows less qualified men to get jobs over potentially higher qualified females, thus not taking full advantage of the available human capital. Understanding the origin of the differences in FLPR among countries is therefore an important first step in reducing the gender gap and correspondingly achieving economic growth. To my knowledge, this is the first study that looks at the development of female labor force participation in former Soviet countries after the fall of the Soviet Union. This paper contributes to the stream of academic literature that investigates the effect of several explanatory variables like economic development, fertility and educational attainment on the female labor participation rate (Alesina, Giuliano and Nunn, 2011; Choudry and Elhorst, 2018; Gaddis and Klasen, 2014; Mammen and Paxson, 2000; Nam, 1991). The data set comprises of 167 countries, including the 15 former Soviet republics and covers the period 1991-2018. By using data on FLPR by the International Labor Organization (ILO) this research finds that after the fall of the Soviet Union, the FLPR in the former Soviet Union countries did become slightly more normal, as the dummy variable for these countries decreased. However, the dummy did not become insignificant. After controlling for several economic and demographic variables, this dummy remained positive and significant. The results of this thesis suggest that there are lasting differences between the former Soviet Union countries and the rest of the world, which cannot be explained by the variables taken into account by this thesis. These variables include economic development, fertility, size of the service sector, educational attainment, government spending and trade openness. It is difficult to determine what drives these vast differences between countries however the results of this thesis hint towards the important role that historical contingencies like ideology can play. Besides ideology, policies possibly also have a direct impact. Policies that enforce egalitarian attitudes on the population may ensure that these gender roles become entrenched in society. Social norms favoring gender equality can be quite persistent and have been shown to lead to higher female labor force participation rates (Antecol, 2008).
4 and will provide details about the data that is used. Chapter 5 discusses the results. Finally, chapter 6 concludes.
2. Theory and literature review
Female labor force participation is a primary indication of the extent to which females participate in the economic activities of any country. FLPR rates have gained considerable interest among researchers and development specialists worldwide due to it being used as a proxy to measure progress being made towards gender equality (Abraham, Ohemeng, and Ohemeng, 2012). There are vast differences in the female labor force participation rate between countries. The determinants of these considerable differences have been well-studied (Alesina, Giuliano and Nunn, 2011). The existing literature on FLPR highlights the importance of several economic and demographic variables in explaining the variation across countries and societies. These variables among other include; per capita income or economic development, specialization in female-friendly sectors, historical contingencies like culture, social norms and/or ideology, fertility, educational attainment, politics.
A large body of literature focusses on economic development in explaining the variation in FLPR (Luci, 2009; Tam, 2010; Gaddis and Klasen, 2014). There are two different approaches explaining the impact of economic development on female labor market participation. The first approach suggests an increase of female labor market participation across all stages of economic growth, this is the ‘modernization neoclassical approach’ based on Becker, 1957 (Luci, 2009). The second approach is based on the work by Boserup (1970), Goldin (1995), and Mammen and Paxson (2002) who recognize a U-shaped relationships between female labor force participation rates and economic development which is usually proxied by GDP per capita (Choudhry and Elhorst, 2018). This approach is called the ‘feminization U’ hypothesis.
The neoclassical approach argues that any sort of discrimination can only be temporary. Discrimination cannot exists in a competitive international environment because it is not consistent with an agent’s optimal behavior that maximizes income or utility. In the long run, all workers will be employed and paid the same wage. According to this approach, if there are persisting inequalities between man and women, this is due to productivity differences, or due to a ‘taste of discrimination’ of employers (Luci, 2009).
5 most women work on farms or in house-hold enterprises they can combine economic activity with childcare. The high levels of female labor force participation in Sub-Saharan African countries agrees with Boserup’s theory showing the importance of women in agriculture in low-income nations (Gaddis and Klasen, 2014).
At the second stage, income starts to grow. There is an increase in urbanization and industrialization, and a decrease in rural sector. The growing demand for labor mobility makes it more difficult for women to combine family and work (Luci, 2009). Men have an easier time finding work because they have privileged access to education and therefore have the ability to adapt more easily to technologies. It is hypothesized that urbanization and industrialization initially decrease the female labor force participation due to structural change and income effect, as the men in the industrial sector now earn enough money to support their entire family. Women encounter restrictions in the labor market due to their childcare responsibilities. This increases their relative labor costs, which leads to employers’ preference for male workers. In addition, during this stage there is generally the prevailing social norm that married women should not seek employment.
During stage three, economic development continues. The exclusion of women from wage activities result in tight labor markets and a rising demand for female workers (Luci, 2009). Furthermore, the educational attainment of women tends to improve during this phase. Cuberes and Teignier (2014) argue that economic growth might lead to reductions in gender inequality in education. Women move back into the labor force as the value of their time increases. Especially, the shift in the composition of production is important. As countries move out of agriculture, and into manufacturing and services, the opportunities for women start to increase (Choudhrey and Elhorst, 2018). The possibility of part-time work, mainly in the service sector, also is important, because this option permits women to combine work outside the household with their domestic activities (Jaumotte, 2003). Ngai and Petrongolo (2013) argue that a rise in the service sector, driven by structural change, results in an increase in the FLPR. Ngai and Petrongolo (2013) argue that women have a natural comparative advantage because the industry is more intensive in ‘brain’ skills and focused on communication and interpersonal skills that cannot easily be automated. They find that expanding the service sector has a positive effect on female employment and wages. Structural changes thus generate a greater demand for female workers.
1980-6 2008. They find some support for the feminization U-curve. In most of their results there is a negative effect of GDP per capita, which is a proxy for economic development, and a positive effect of its squared term. However, their assessment is that support for the feminization-U curve relies mostly on the data source that is used. Therefore, their conclusion is that the feminization-U hypothesis has some explanatory power but this is not robust. Furthermore, the hypothesis fails to explain certain substantial differences in FLPR among different countries. Choudry and Elhorst (2018) argue that economic development alone is insufficient in explaining the large differences in female labor force participation rates. Next to the feminization-U hypothesis there are other sources of literature that try to explain the considerable differences in female labor force participation between countries. Some authors have presented explanations emphasizing the role of historical contingencies. Multiple studies find a considerable influence of culture and/or social norms. Social norms may limit the ability of women to accept paid employment, especially in manual jobs (Mammen and Paxson, 2002). Goldin (1955) argues that certain societies stigmatize the husband of women who do blue-collar work because this goes against prevailing social norms that a man must provide for his family (also see Boserup, 1970). More recent studies support the theory put forth by Goldin and Boserup. More traditional societies in the Middle East, consider women mainly as housewives and mothers as opposed to breadwinners (Olmested, 2011). This is underscored by a study carried out by Gaddis and Klasen (2014). They find particularly large negative fixed effects for countries in the Middle East and Latin America where prevailing social norms limit the entry of women into the labor market. Perceptions of women’s role as homemakers have been proven to be quite persistent over time (Fortin, 2005). Araújo and Scalon (2005) present evidence of the effect of cultural characteristics of Brazil on the FLPR and argue that traditional attitudes towards gender role have become entrenched in society and therefore can be long lasting. Another example on how persistent culture, social norms and gender biases can be is given by Alesina, Giuliano and Nunn (2011). They argue that the current division of labor is generated by historical differences in agricultural systems. They find that societies that traditionally practiced plow agriculture have significantly lower FPLR today. This is because plow agriculture requires a lot of strength, favoring men over women. Societies that traditionally used hand-held tools for cultivation are more likely to have higher FLPR due to the fact that women actively participated in this kind of agriculture. Antecol (2008) looks at FLPR in countries in Europe, the Middle East, and Asia and includes survey results of men’s attitude towards the distribution of roles among different genders. She finds that women are more likely to participate in the labor market if their male partners are less conservative considering gender roles.
7 positive effects on the female labor force participation rate and falling fertility rates can contribute significantly to a takeoff in economic growth. Reducing the fertility rate by changes in abortion laws also has been shown to lead to higher female labor participation rates. Legalization of abortion in the United States led to an increase of the FLPR of black women (Angrist and Evans, 2000). Changes in legislation have often been used as an instrument for fertility, finding a negative effect of fertility on female labor force participation (Bailey, 2006). Other sources on the other hand, argue that females will use labor market opportunities to seek employment, not only in low-income but also high-income countries. In their view, children are one of the main reason for women to participate on the labor market, because they need to provide for their families (Abraham, Ohemeng, and Ohemeng, 2012).
Educational attainment is another important factor in explaining the difference in female labor force participation rates. There are good reasons to assume that educational attainment is positively related to intelligence, ambition, and that these factors in turn are related to labor force participation (Bowen and Finegan, 1966). When educational attainment starts to improve, so does the value of a woman’s time outside the household (Choudry and Elhorst, 2018). On the one hand, education can be seen as an investment to increase future earning’s potential and therefore education incentives people to seek employment. Nam (1991) adds that more educated women tend to be more active in the labor market due to the fact that education increases income aspirations. On the other hand, increasing the level of education can also acts as a deterrent to female labor market entry. An increase in the amount of years that females enjoy education can delay their entry unto the labor market. This is because women cannot be enrolled in formal education and work at the same time (ILO, 2018). However, it is hypothesized that in the long-run education has a positive effect on the FLPR. Most empirical efforts confirm this hypothesis and find that education is positively related to FLPR whereas fertility negatively affects the FLPR (Choudry and Elhorst, 2018; Contreras and Plaza, 2010; Nam, 1991).
Aside from the variables mentioned above, globalization and trade liberalization also need to be addressed. Gaddis and Pieters (2012) study the effect on trade liberalization in the late 1980’s on the female labor force participation in Brazil and find that trade reforms can have profound effects on the labor market. This is because trade liberalization is expected to lead to greater economic activity. Their results show that declining trade protection led to an increase of the female labor force participation in Brazil. However, these results must be interpreted with caution as the effects of trade liberalization depend on the countries’ initial factors of endowments. If a trade liberalization leads to expansion in a male-labor intensive tasks, this could depress the female labor force participation rate. If the emerging sector creates requires female-labor intensive tasks, this could lead to higher economic activity by females (Catagy and Berik, 1990; Rees and Riezman, 2013). Cooray, Dutta and Mallick (2017) find an overall positive effect of increased trade openness in Africa on both the female and male labor force participation rates.
8 Sweden and Great Britain to identify the effect that different family oriented policies have on the FLPR. They find that in Sweden, where the FLPR is one of the highest in the world, several provisions are in place enabling women to combine work and a family. An improved and equal status in the job market encourages women to participate in the labor market (Mitra, 2005). In Great Britain, which relies more on the market rather than state sponsored benefits, the FLPR is much lower. Alongside policies, Gustafson et al. (1996) also underscore role of human capital, as they find that when the human capital of women is sufficiently high, women are more likely to combine a family with work regardless of the family-oriented policies. Expansionary fiscal policy has also proven to increase the labor force participation rate. Increases in government spending have a positive effect on the demand for goods and services which can increase the labor force participation and the employment rate (Brücker and Pappa, 2011). Additionally, ideology can have a strong influence on female labor participation rates. Labor shortages combined with an ideology that promotes equality can lead to remarkably high FLPR (Gaddis and Klasen, 2014). The role that a combination of ideology and policies can play in promoting high female labor force participation rates is highlighted in the next chapter.
3. Empirical background
Both ideology and policy can play an important determinant of female labor participation. The Soviet Union was based on certain fundamental principles that were laid out by Karl Marx and Friedrich Engels. Karl Marx liked to quote the utopian socialist Fourier who said: ‘’The change in an historical era can always be determined by the progress of women towards freedom, for the victory of human nature is most evident in the relation of women to men, of the weak to the strong. The degree of the emancipation of women is the natural measure of gender emancipation’’ (Farnsworth, 2019; Uska, 2005). Marx and Engels believed that there where capitalism exists, men remain to more privileged than women. Amongst other things, their focus was on facilitating the empowerment of women and making sure that the participation of women in economic, political and social life was the equal to that of men (Uska, 2005). Full economic independence could be the only base for full gender equality. The believes that Marx and Engels espoused were later pledged in the 1936 constitution of the USSR stating that ‘’women in the USSR are accorded equal rights with men in all spheres of economic, state, cultural, social, and political life.’’ To what extent this was upheld is under debate, but the fact remains that gender equality is a key pillar on which the Soviet Union was built.
9 female labor participation rates that were amongst the highest in the world (Brainerd, 2000). Part-time employment was unavailable in these countries, so women generally chose employment in the health and education sector because here they had shorter and more flexible hours which allowed them to combine the work with childcare.
By the beginning of the 1990’s, the Soviet Union lost its cohesion during the tenure of Mikhail Gorbachev. This caused an increase in secessionist movements and by 1991, there were major market reforms on the way in almost all Soviet Countries. These reforms include wage and price liberalizations, trade liberalization, privatization of state-owned enterprises, and tax and legal reforms. The Eastern European countries were fast and already in 1991, they created institutions necessary for a market economy. Other countries, like the Central Asian nations were a bit slower but eventually followed. In all countries, the fall of the Soviet Union and the reforms that followed initially led to a decrease in GDP, as is evident from figure 1, and an increase in inflation. Labor market institutions have since changed dramatically, and differ substantially per country (Brainerd, 2000). This could potentially have profound effects on the female labor participation rates in the respective countries. The following section of this thesis looks at how labor participation rates have developed in the 15 former Soviet republics since the dissolution of the Soviet Empire.
4. Data and methodology 4.1 Data source and description
This study is based on data taken from the International Labor Organization database (2019). Female labor force participation the variable of interest in this study and is defined as the proportion of the population of age 15-64 that is ‘economically active’ divided by the total female population of the same age group. This data is available for the years 1990-2018. Labor force participation is a good measure to evaluate the working-age population in an economy, as it reflects the number of people who are interested in participating in the workforce. Throughout this thesis, labor force participation is defined as being economically active. Women who are economically active are either employed or self-employed meaning that they supply labor for the production of economic goods and services. Women who are unemployed but are still actively looking for work are also included in the female labor participation rate. The notion of economically active is an important concept as not everybody who works is included. Some women invest their time in activities that are not generally seen as ‘market labor’. These includes activities like raising children and house-hold chores like cooking and cleaning. This is not treated as economically active because these activities do not translate into income. People who are not looking for a job and people who are not interested in finding one, also do not count as economically active (ILO, 2019).
10 do report some sort of labor market information based on household surveys for example. Therefore, for most countries the ILO estimates are a combination of the data sources listed above. The ILO uses several econometric models to produce estimates of labor market indicators in the countries and the years where there is no data available. The purpose of this is to establish a balanced panel data (ILO, 2019).
Data on GDP per capita is taken from the Penn World Tables. The most recent available version of the PWT, version 9.1 reports GDP at chained PPP in millions of 2011 US$. This is divided by the population of the respective country which is also taken from the PWT 9.1. GDP per capita is used in this thesis as a proxy indicator for economic development, as is common in this line of research (Gaddis and Klasen, 2014). Figure 1 illustrates the fact that GDP per capita increased during the sample period most of the analyzed countries. The Soviet Union was a tightly integrated economic space. Following the fall of the Soviet Union, all 15 countries initially experienced negative economic growth, as seen by the decrease in GDP per capita. After the 2000s, economic growth took off in most countries, to such an extent that GDP per capita started to exceed pre-Soviet levels. However, there are three countries in the sample where the level of GDP per capita in 2018 does not exceed the level of GDP per capita in 1990. These countries are Tajikistan, Ukraine, and Kyrgyzstan.
Figure 1: GDP per capita, PPP adjusted, 1990-2017
Source: Penn World Tables version 9.1, 2019. Own computation.
The earliest data available on labor force participation rates is 1990. Table 1 shows that the female labor force participation rate in the analyzed countries before 1991 is considerably larger than the world average. The difference between the former Soviet Union countries and the rest of the world is 5.1 percentage points. This difference has become much smaller in 2018 and was then only 2.4 percentage points. Figure 2 shows the development of the FLPR in each country separately. The individual experience the former Soviet countries had with the female labor participation rates differs to quite some extent. Table 2 shows the difference between 1990 and 2018 in percentage points. Some countries experienced large increases in the FLPR.
11 Azerbaijan for example, has seen an increase of 10.7 percentage points. Other countries like Belarus, Turkmenistan and Uzbekistan also experienced an increase, albeit smaller. Ukraine, Moldova and Russia are examples of countries where the female labor force participation rate decreased, in some cases with more than 3 percentage points. Interestingly enough, the countries that experienced a decrease in GDP in the period 1990-2018 follow somewhat the same trajectory with regards to FLPR. Ukraine and Kyrgyzstan both saw a decrease in their FLPR in the period 1990-2018 whereas the FLPR in Tajikistan remained exactly the same. Figure 2: The development of the FLPR in the former Soviet countries
Source: International Labor Organization, 2019.
Table 2: Changes in the FLPR in the former Soviet countries in percentage points
Country FLPR 1990 FLPR 2018 ΔFLPR 1990-2018 Armenia 49,0 51,6 2,7 Moldova 47,3 38,9 -8,4 Estonia 54,9 56,2 1,3 Latvia 53,4 55,0 1,6 Lithuania 54,6 55,9 1,4 Georgia 58,1 58,0 -0,2 Azerbaijan 52,1 62,8 10,7 Tajikistan 45,4 45,4 0,0 Kyrgyzstan 50,8 48,0 -2,8 Belarus 53,2 58,2 5,0 Uzbekistan 50,0 53,8 3,8 Turkmenistan 49,9 53,3 3,4 38 43 48 53 58 63 68 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Fe m al e Lab or F or ce P ar tic ip at ion R at e ( % )
Armenia Moldova Estonia Latvia Lithuania
Georgia Azerbaijan Tajikistan Kyrgyz Republic Belarus
12
Ukraine 54,8 46,6 -8,1
Kazakhstan 65,3 65,2 0,0
Russia 59,3 56,3 -3,0
Source: International Labor Organization, 2019.
4.1.1 Control Variables
Studies that use GDP per capita, as a proxy for economic development, often find that its coefficient is negative and that the squared term is positive. Choudhry and Elhorst (2018) argue that only including the squared term of GDP is not significant to capture the U-shaped relationship between female labor force participation and economic development but instead they claim that it is necessary to investigate whether economic development impacts the FLPR in combination with other explanatory variables. As is addressed in the literature review, there are multiple variables that have been proven to affect the female labor force participation. Some relatively recent studies like the one carried out by Gaddis and Klasen (2014), Luci (2009) and Tam (2011) fail to include certain important control variables and therefore suffer from omitted variable bias (Choudhry and Elhorst, 2018). The following explanatory variables have been taken into consideration in this thesis; fertility, the share of the population employed in the service sector, educational attainment, government expenditure as a percentage of GDP and trade openness.
13 causing higher female labor force participation rates (Cooray, Dutta and Mallick, 2017; Gaddis and Pieters, 2012). Table 3 provides the summary statistics of the compiled data set. Table 9 in appendix A shows the correlation matrix. There is quite a strong positive correlation between GDP per capita and the size of the service of the service sector. This is understandable as countries that specialize in services generally are further along the process of structural change and therefore have higher GDP per capita. The correlation between educational attainment and GDP per capita is also fairly high. Again, this is coherent as many studies show that schooling is positively correlated with GDP per capita (Barro, 1991; Bils and Klenow, 1998).
Table 3: Summary statistics
Variable Obs. Mean Std. Dev. Min Max
FLPR (%) 6,757 50.0333 15.5689 5.8230 90.7840
lnGDP 5,066 8.9726 0.2303 5.4076 11.9412
lnGDP2 5,066 82.0599 22.1197 29.2421 142.5919
Size of the service
industry (%) 6,524 47.6823 20.1766 5.0690 87.5910 Fertility (amount of births) 6,626 3.2000 1.6520 0.827 8.606 Educational attainment 4,390 11.6630 3.8458 0.6 23.3 Government expenditure (%) 5,616 16.0501 6.7602 0.9112 135.794 Trade openness 6,048 82.0449 51.5463 0.0209 860.8
4.1.2 Drawbacks of the data
There are some drawbacks of the data that are worth mentioning. Firstly, a large part of the theoretical framework is based on structural change. As countries experience structural change, they move from low productivity agricultural activities to sectors characterized by a higher productivity like the manufacturing and the service sector. The GGDC 10-sector Database provides a dataset on sectoral productivity in African, Asian, and Latin-American countries. However, there is no sectoral productivity data available for the 15 countries that are analyzed in this thesis. Therefore, this thesis looks at an alternative source of data, the size of the service sector. As it is assumed that more developed economies are further in the process of structural change and consequently, these countries have a relatively larger share of their working population employed in the service sector. This assumption is supported by the literature (Choudry and Elhorst, 2018; Gaddis and Pieters, 2012; Jaumotte, 2003; Ngai and Petrongolo, 2013).
14 and what does not. To underscore this argument, Tzannatos (1999) argues that differences in the female labor participation rates can be nothing more than a statistical artefact due to fact that definitions change over time, and they differ between countries. For example, some countries only measure persons employed in paid employment whereas other countries include all persons that are engaged. Meaning that alongside paid employment they also take into account people who receive some sort of remuneration. Another example is that measurement of total employment sometimes include a limited amount of hours worked. However, this measurement differs per country. Measurement has also proven to be difficult due to frequency of data collections. An annual average based on 12 months of observations is most likely to differ from an average based on a few observations making it not directly comparable (ILO, 2019). The ILO recognizes that data from countries with limited labor market information could be inaccurate, and should therefore be interpreted with caution.
4.2 Methodology
4.2.1 Difference in FLPR between Soviet countries and the rest of the world in 1990 The starting point is to examine whether the female labor force participation rates in the analyzed countries differ significantly during Soviet times from the rest of the world. Unfortunately, there is only one year of data available before 1991. For that reason, only 1990 is taken into consideration. Equation (1) tests whether the FLPR in the former Soviet countries is significantly different from the FLPR of the rest of the world in 1990.
𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝑖𝑖 = 𝑎𝑎 + 𝛿𝛿(𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑖𝑖)+ 𝜇𝜇𝑖𝑖 (1)
𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝑖𝑖 represents the female labor force participation that varies across countries (i) in 1990.
A Soviet dummy variable is included to determine whether there is in fact a significant difference between Soviet countries and countries in the rest of the world. This dummy takes on the value 1 if the country was incorporated in the Soviet Union in 1990, and the value 0 if otherwise.
4.2.2. Female labor participation rates in the former Soviet Countries
15 consideration. First, there is the unconditional model. This model is essentially the same as equation (1). The only difference is that now the period of interest is after the fall of the Soviet Union instead of before. Equation (1) only captures the year 1990, whereas the following model takes into consideration every year after the fall of the Soviet Union for which data is available. The effect of being a former Soviet republic on the female labor participation rate is estimated in the following regression:
𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝑖𝑖𝑖𝑖 = 𝑎𝑎 + 𝛿𝛿(𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑖𝑖)+ 𝜇𝜇𝑖𝑖𝑖𝑖 (unconditional)
This model is called the unconditional model and only looks at the effect that being a former Soviet Union country has on the female labor participation rate. In this equation, 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝑖𝑖𝑖𝑖
represents the female labor force participation that varies across countries (i) in year (t). The effect of being a former Soviet country is captured by δ.
As a benchmark control, economic development is included. This model is therefore called the benchmark model and is estimated by the following equation:
𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝑖𝑖𝑖𝑖 = 𝑎𝑎 + 𝛿𝛿(𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑖𝑖) + 𝛽𝛽 (ln 𝑔𝑔𝑑𝑑𝑔𝑔)𝑖𝑖𝑖𝑖 + 𝛾𝛾 (ln 𝑔𝑔𝑑𝑑𝑔𝑔) 𝑖𝑖𝑖𝑖2 + 𝜇𝜇𝑖𝑖𝑖𝑖 (benchmark)
This model includes both GDP per capita (PPP deflated) and its squared term. GDP per capita is transformed into the natural logarithm of GDP per capita in order to correct for the skewedness of the data. The squared term of GDP per capita is included to test the validity of the feminization-U hypothesis. Based on the assumptions of this hypothesis, as laid out in chapter 2, it is expected that β<0 and γ>0. Meaning, that an increase in GDP initially leads to a decrease in FLPR but after reaching a certain threshold, the additional effect of GDP on the FLPR becomes positive. Again, a Soviet dummy is included.
Choudhry and Elhorst (2018) argue that only looking at GDP per capita is insufficient and that it is necessary to take certain explanatory variables into account. The following model is based on the benchmark models but adds additional explanatory variables. This third model is called the augmented model and is represented by the following equation:
𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝑖𝑖𝑖𝑖 = 𝑎𝑎 + 𝛿𝛿(𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑖𝑖) + 𝛽𝛽 (ln 𝑔𝑔𝑑𝑑𝑔𝑔)𝑖𝑖𝑖𝑖+ 𝛾𝛾 (ln 𝑔𝑔𝑑𝑑𝑔𝑔) 𝑖𝑖𝑖𝑖2+ 𝜌𝜌𝑋𝑋𝑖𝑖𝑖𝑖
+ 𝜇𝜇𝑖𝑖𝑖𝑖 (augmented)
Equation (4) includes δ, β and γ but adds ρ. This coefficient captures the effect of the controlling variables (Xit) that have been outlined in section 3.1.1. These variables include fertility,
educational attainment, the size of the service sector, government expenditure and trade openness.
16 GDP. The same control variables as in the equation above have been included. This model is referred to as the control model.
𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝑖𝑖𝑖𝑖 = 𝑎𝑎 + 𝛿𝛿(𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑖𝑖) + 𝛽𝛽 (ln 𝑔𝑔𝑑𝑑𝑔𝑔)𝑖𝑖𝑖𝑖 + 𝜌𝜌𝑋𝑋𝑖𝑖𝑖𝑖+ 𝜇𝜇𝑖𝑖𝑖𝑖 (control)
5. Empirical results
The regression results of equation (1) are represented in Table 4. The coefficient of the dummy variable has a value of 5.16 so the female labor participation rate was higher in the Soviet countries in 1990. This coefficient is significant at the 5% level. This means that the FLPR were indeed significantly higher in the countries that were incorporated in the Soviet Union, than the FLPR were in the rest of the world (in 1990). The rest of this thesis will analyze how the female labor force participation rates have developed in the former Soviet republics since the fall of the Union in 1991.
Table 4: The difference in FLPR between former Soviet countries and the rest of the world, in 1990
Variable (1) Soviet dummy 5.1621 (2.5226)** Constant 48.1053 (0.6400)*** Observations 233 ρ-value 0.0411 R2 0.0060
Standard errors in parentheses.
∗∗∗p < 0.01, ∗∗p < 0.05, ∗p < 0.1.
Table 5 demonstrates the results of the unconditional model, the benchmark model, the augmented model and the control model. Column 1 shows the result of the unconditional model. It can be observed that the Soviet dummy coefficient is positive and significant at the 1% level, which means that Soviet countries have significantly higher FLPR throughout the period 1991-2018. Column 2 adds the log of GDP per capita and the log of GDP per capita squared, from here on out referred to as GDP and GDP squared. The coefficient for GDP is negative and significant and that of GDP squared is positive and significant. These results are consistent with the feminization U-hypothesis, meaning that economic development initially has a negative effect on the FLPR but after a certain threshold is reached, increased economic development starts having a positive effect on the FLPR. This means that women benefit to some extent from economic development but only after a certain point. Interestingly enough, the coefficient of the Soviet dummy in column 2 increases when adding GDP and GDP squared, meaning that economic development as a factor alone, cannot account for much of the variation in FLPR between the former Soviet Union countries and the rest of the world.
17 government expenditure and trade openness decreases the value of the dummy coefficient. This signals that the factors identified in the literature are important in explaining some of the considerable differences in the FLPR. As expected, fertility is negative and significant, albeit on the 10% level, which means that increases in the amount of children per woman depresses the female labor participation rate. The size of the service industry is significant but negative, this is opposite to what we would expect. The theory argues that women have a comparative advantage in the service sector. Therefore, it is expected that a country with a larger service sector has a higher FLPR. However, this hypothesis is not supported by the results of this thesis. This suggests that growth in the service sector does not increase the employment opportunities for women. Educational attainment is significant and positive, meaning that increasing the years of education increases the FLPR. Presumably because education increases the value of women’s time as well as their incentives to seek employment. Furthermore, the coefficients for both GDP and GDP have become larger after controlling for the factors mentioned above. So conditional on the controlling variables, GDP and GDP squared capture more of the variance between FLPR. In the augmented model, the value of the soviet dummy’s coefficient is still positive and significant meaning that the explanatory variables that are included are not sufficient in explaining the variation between the FLPR in the former Soviet countries and the rest of the world.
Table 5: OLS estimates, period 1991-2018
Dependent variable: female labor force participation rate
Variable Unconditional Benchmark Augmented Control
Soviet dummy 3.5303 (0.3300)*** (0.4457)*** 6.6170 (0.5555)*** 3.7677 (0.5295)*** 2.7104 lnGDP per capita - -64.6353 (2.7266)*** (3.5691)*** -72.7311 (0.4546) 0.0901 lnGDP per capita2 - 3.4994 (0.1517)*** (0.1861)*** 3.9379 - Fertility - - -0.5039 (0.3016)* (0.2604)*** 1.9856 Size of the service
industry (in %) - - (0.0286)*** -0.2961 (0.0281)*** -0.3961 Educational attainment - - 0.1466 (0.1118)*** (0.1228)*** 1.6972 Government expenditure (% of GDP) - - (0.0449) -0.0577 (0.0483) -0.0527 Trade openness - - -0.0031 (0.0040) (0.0040)*** 0.0246 Constant 49.8685 (0.2044)*** (12.0644)*** 342.6589 (17.4694)*** 379.4281 (3.7954)*** 41.9387 Observations 6,524 4,509 3,823 3,823 ρ-value 0.0000 0.0000 0.0000 0.0000 R-squared 0.0031 0.1788 0.2617 0.1406
18 One may argue that the feminization-U hypothesis is not relevant in this particular study as most of the countries included in the sample were already quite industrialized after 1991. Therefore, the control model tests the feminization U-hypothesis. This specification does not include the squared term of GDP per capita, and tests whether the relationship is purely linear. The results can be found in column 5 of Table 5. It is expected that once you remove the non-linear component of GDP per capita, the coefficient becomes positive. The hypothesis is that once a country experiences a growth in terms of GDP, the female labor force participation rate will increase as well. The relationship between FLPR and economic development is positive, however is it is not significant. This is contrary to what the literature suggests. Surprisingly enough, the coefficient for fertility also changes sign. With the exclusion of GDP squared, this coefficient becomes positive and significant meaning that higher fertility leads to increased economic participation by women. This is interesting as this is contrary to what the theory predicts. A possible explanation could be that women seek labor market opportunities in order to provide for their families. The control model has a lower R2 than the benchmark and the
augmented model. Furthermore, the theory suggests that economic development is an important determinant of FLPR but in this case it is insignificant. It is therefore reasoned that the control model is not a particularly sound model.
This study looks at the development of the female labor participation rates in former Soviet countries. The results of table 5 show that economic development plays a somewhat important role in capturing the variation in female labor participation rates after the fall of the Soviet Union. The inclusion of GDP and its squared term, as a proxy for economic development, capture some of the variation in FLPR. However, these are not the only explanatory variables of importance. The augmented model shows that adding the additional explanatory variables captures more of the variation and decreases the coefficient of the Soviet dummy variable in comparison with the benchmark model. The augmented model also has the highest R-squared, and is therefore assumed to be the most valid model in the analysis.
As is argued before, the Soviet Union was a tightly integrated economic space. Countries needed time to adjust to the dissolution of the Soviet empire. Figure 1 shows that all countries initially experienced a decrease in their GDP per capita and in some countries the current level of GDP does not exceed their Soviet levels. Furthermore, some countries were quick to adopt the institutions necessary for a market economy whereas others acted more slowly. It could therefore be argued that the analyzed time-period is too large and that it makes sense to look at smaller periods of time. The following part examines the unconditional model, the benchmark model and the augmented model further. Measuring the extent of gender differences in the labor market and their change over time can provide useful information. It is analyzed whether different time periods have an influence on the value of the coefficients. Four different time periods are taken into account with 7 year intervals namely 1991-1997, 1998-2014, 2005-2011 and 2012-2018.
Table 6: OLS estimates of the unconditional model, 7 year periods
Dependent variable: female labor force participation rate
Variable 1991-1997
19 Soviet dummy 5.0275 (0.6162)*** (0.567)*** 3.7237 (0.6798)*** 2.7236 (0.7362)*** 2.6464 Constant 48.44 49.4365 50.5216 51.0731 (0.4246)*** (0.4113)*** (0.4012)*** (0.3952)*** Observations 1,631 1,631 1,631 1.631 ρ-value 0.0000 0.0000 0.0000 0.0000 R-squared 0.0059 0.0034 0.0019 0.0019
Robust standard errors in parentheses. ∗∗∗p < 0.01, ∗∗p < 0.05, ∗p < 0.1.
Table 6 shows the result of the unconditional model analyzed over time. Without the inclusion of any of the explanatory variables, the coefficient of the Soviet dummy remains significant and positive however the value of the dummy decreases considerably over time. The value of the Soviet dummy starts with 5.03 in the first period and drops to 2.65 in the last period. This is a decrease of more than 50%. These results correspond with table 1, that shows that the difference between the FSU countries and the rest of the world has become a lot smaller and in fact decreased with more than 50%. However, this is not a particularly good model as it does not include any of the explanatory variables that have been shown to impact the FLPR. Furthermore, the R-squared is much lower than the benchmark and augmented model.
Table 7: OLS estimates of the benchmark model, 7 year periods
Dependent variable: female labor force participation rate
Variable 1991-1997
(benchmark) (benchmark) 1998-2004 (benchmark) 2005-2011 (benchmark) 2012-2018 Soviet dummy 9.0639 (0.8406)*** (0.9219)*** 6.934 (0.9496)*** 6.0053 (0.9872)*** 5.5086 lnGDP per capita -59.3285 (6.7812)*** (5.0547)*** -74.1659 (4.2981)*** -71.4162 (4.5648)*** -70.3293 lnGDP per capita2 3.2547 (0.3890)*** (0.2885)*** 4.0534 (0.2401)*** 3.8145 (0.2501)*** 3.7282 Constant 312.865 (29.1541)*** (21.8022)*** 381.8985 (18.8080)*** 378.9689 (20.4546)*** 377.9846 Observations 1,169 1,169 1,169 1,002 ρ-value 0.0000 0.0000 0.0000 0.0000 R-squared 0.1614 0.2215 0.2166 0.1860
Robust standard errors in parentheses. ∗∗∗p < 0.01, ∗∗p < 0.05, ∗p < 0.1.
20 but remain higher than their initial value. The results of Table 7 show that the female labor force rate in the FSU countries became slightly more normal, when controlling for economic development, but they remained significantly higher than the FLPR in the rest of the world.
Table 8: OLS estimates of the augmented model, 7 year periods
Dependent variable: female labor force participation rate
Variable 1991-1997
(augmented) (augmented) 1998-2004 (augmented) 2005-2011 (augmented) 2012-2018 Soviet dummy 6.3745 (1.2845)*** (1.2954)*** 3.5589 (1.0107)*** 3.8109 (1.1587)*** 4.2417 lnGDP per capita -86.4548 (6.9102)*** (7.3716)*** -80.4620 (6.5141)*** -70.4165 (7.8431)*** -70.7429 lnGDP per capita2 5.0358 (0.3659)*** (0.3834)*** 4.4396 (0.3362)*** 3.6981 (0.4005)*** 3.5953 Fertility -0.1421 (0.5165) (0.6053) -0.8335 (0.6436) 0.2871 (0.7799) 0.2809 Size of the service
industry (0.0528)*** -0.5708 (0.0553)*** -0.3381 (0.0500)*** -0.1521 (0.0572) 0.0219 Educational attainment (0.2134)*** 1.4571 (0.2170)*** 1.3084 (0.2414)*** 1.6596 (0.2545)*** 1.1261 Government expenditure (0.0666)* 0.1274 (0.0992) -0.0945 (0.0992)*** -0.3379 (0.1038)*** -0.2751 Trade openness 0.0077 (0.0087) (0.0087) -0.0017 (0.0072) -0.0030 (0.0073) 0.0058 Constant 421.2829 (31.8359)*** (35.8561)*** 413.1001 (32.5315)*** 371.7912 (39.6835)*** 381.0492 Observations 946 1,030 1,070 777 ρ-value 0.0000 0.0000 0.0000 0.0000 R-squared 0.3505 0.3026 0.2563 0.2322
Robust standard errors in parentheses. ∗∗∗p < 0.01, ∗∗p < 0.05, ∗p < 0.1.
21 Table 5 shows that fertility has a negative coefficient in the augmented model in the overall sample. It is hypothesizes that childbirth has a negative effect on the FLPR. This is because the more children women have, the more difficult is becomes to combine raising children with a career. Table 8 however shows, that this negative effect is inconsistent. With the inclusion of other controlling variables, this impact of fertility on the FLPR is insignificant in all four sub-periods. The size of the service industry is surprisingly negative throughout the sample and becomes insignificant in the last period. For the whole of the sample, educational attainment remains positive and significant. This is consistent with the literature and shows that increasing the amount of years that women enjoy education encourages their entry into the labor market. The findings of this thesis suggest that the relationship between FLPR and the explanatory variables that have been taken into account is a complex one and varies over time. Nevertheless, the Soviet dummy remains significant and positive in all the specifications in this study. The FLPR in the former Soviet countries has become slightly more normal as the dummy variable decreased from its value in the first period. However, the Soviet dummy remains significant and positive in all the specifications in this study. This indicates that almost 30 years after the fall of the Soviet Union, there are lasting differences between the former Soviet Union countries and the rest of the world. It is difficult to determine what drives these lasting differences in the FSU countries. A possible explanation could be rooted in the gender equality promoting policies and ideology of the socialist past.
5.1 Robustness
Gaddis and Klasen (2014) find that support for the feminization-U hypothesis and the role that economic development plays in explaining FLPR depends on the data source that is used. In order to provide a robustness check, this thesis also takes into account another source of GDP per capita. This data on GDP per capita is taken from the World Bank’s Development Indicators (2019). The World Bank’s Development Indicators report GDP at constant 2010 US$. Table 10 in appendix B reports the results for the augmented model. The dummy variable remains positive and significant, although in some periods on the 5% level, when using a different source of GDP. The value of the coefficient for the Soviet dummy is quite a lot smaller than the estimations in Table 8. Using a different source of GDP per capita more than halves the coefficient of the FLPR, going from 6.4 to 2.9 in the robustness check. Contrary to the results of the augmented model in Table 8, the coefficient of the dummy variable does not decrease after the first period but only increases. The results of Table 10 do confirm that there are lasting differences in the female labor force participation rates between the FSU countries and the rest of the world, even after controlling for certain demographic and economic variables. The robustness check also confirms that economic development is a significant variable in explaining some of the variation in the FLPR, since β<0 and γ>0. Educational attainment also remains positive and significant, underscoring the importance of education in encouraging increased female labor force participation rates.
6. Conclusion
22 Decreases in population growth may negatively affect economic growth. Especially in ageing societies, policies aimed at increasing the female labor force participation could reduce the strain on a tight labor market. The aim of this thesis is to analyze the development of the female labor participation in former Soviet Countries. There is a stream of literature that focusses on the determinants of female labor force participation but to my knowledge the former Soviet republics have never been scrutinized.
The factors that potentially explain the large variation in female labor force participation rates between countries have been identified as economic development, proxied by GDP per capita and GDP squared, fertility, education, the size of the service sector, government expenditure and trade openness. The existence of a non-linear relationship between economic development and female labor participation rates imply that there is a tradeoff to be made between economic growth and gender equality, recall that β<0 and γ>0. However, economic development is far from the only factor impacting the female labor force participation rates. The results of this thesis show that aside from economic development, educational attainment also has an impact on the FLPR. Increasing the educational attainment increases female participation in economic life.
The results of this thesis show that after controlling for several economic and demographic variables, the Soviet dummy continues to be significant and positive. These results hint towards the importance of historical contingencies like ideology and social norms about the role of women in society in explaining the large differences in FLPR. Gender equality promoting polices possibly have a lasting effect on the female labor force participation rate. Women who hold more egalitarian attitudes about gendered family roles will be more likely to enter the paid labor force than women who are more supportive of gender-differentiation in families (Cunningham, 2008). Enforcing egalitarian attitudes on the population may ensure that these gender roles become entrenched in society. This results however, need to be interpreted with caution as the ILO recognizes that the data coming from developing countries might suffer from a lack of comparability. Furthermore, the list of controlling variables might be inexhaustive, causing there to be an omitted variable bias.
This research contributes to a number of studies that have provided evidence that the vast differences in FLPR may be explained by differences in ideology and beliefs about the appropriate role of women in society (Alesina, Giuliano and Nunn, 2011; Fortin, 2015). Further research needs to be done about the lasting role of ideology on female labor force participation but the results of this thesis hint towards the fact that ideology and policies aimed at gender equality can have a positive and long lasting effect on women’s labor force participation rates, even almost 30 years after the transition from socialism to capitalism. Considering the positive relationship between FLPR and economic development, active measures to enhance women’s economic empowerment should be a central to policies aimed at reducing poverty and inequality.
23 References
Abraham, A.Y., Ohemeng, F.N.A., Ohemeng, W. (2012). Female labour force participation: evidence from Ghana. International Journal of Social Economics, 44(11): 1489-1505. Amate-Fortes, I., Guarnido-Rueda, A., and Molina-Morales, A. (2015). On the perpetuation
of the Situation of Economic and Social Underdevelopment in Africa. Journal of African American Studies, 19: 233-263.
Angrist, J., and Evans, W. (2000). Schooling and labor market consequences of the 1970 state abortion reforms. Research in Labor Economics, 18: 75-113.
Antecol, H. (2003). Why is there Cross-Country Variation in Female Labor Force Participation Rates? The Role of Male Attitudes towards Family and Sex Roles. Working Paper No.03,03. Claremont, CA: Claremont McKenna College.
Alesina, A.F., Giuliano, P., and Nunn, N. (2011). On the origins of Gender Roles: Women and the Plough. NBER Working Papers No.17098. Cambridge, MA: National Bureau of Economic Research.
Austen, G. (2005). Labour, land and capital in Ghana: From slavery to free labour in Asante, 1807-1956. Rochester: Rochester University Press.
Araújo, C., and Scalon, C. (2005). Gender, family and work in Brasil. Editora da fundacao getulio vargas, 1: 15-76.
Bailey, M.J. (2006). More power to the pill: The impact of contraceptive freedom on
women’s lifecycle labor supply. The Quarterly Journal of Economics, 121(1): 289-320. Barro, R.J. (1991). Economic Growth in a Cross Section of Countries. Quarterly Journal of
Economics, 106(2): 407-444.
Bils, M., and Klenow, P.J. (1998). Does Schooling Cause Growth or the Other Way Around? NBER Working Paper No. 6393. Cambridge, MA: National Bureau of Economic Research.
Bloom, D.E., Canning, D., Fink, G., and Finlay, J.E. (2009). Fertility, female labor force participation, and the demographic dividend. Journal of Economic Growth, 14: 79-101. Boserup, E. (1970). Women’s role in economic development. New York, NY: St. Martin. Bowen, W.G., and Finegan, A. (1966). Educational Attainment and Labor Force Participation.
The American Economic Review, 56(1/2): 567-582.
Brainerd, E. (2002). Women in Transition: Changes in Gender Wage Differentials In Eastern Europe and the Former Soviet. Female Industrial and Labor Relations Review, 54(1): 138-162.
Brücker, M., and Pappa, E. (2012). Fiscal Expansions, Unemployment and Labor Force Participation: Theory and Evidence. International Economic Review, 53(4): 1205-1228. Cagatay, N.,, and Berik, G. (1990). Transition to Export led Growth in Turkey: Is there a
24 Choudhry, M.T., and Elhorst, P. (2018). Female labour force participation and economic
development. International Journal of Manpower, 39 (7): 896-912.
Cooray, A., Dutta, N., and Mallick, S. (2017). Trade Openness and Labor Force Participation in Africa: The Role of Political Institutions. Industrial Relations: A Journal of Economy and Society, 56(2): 319-350.
Contreras, D., and Plaza, G. (2010). Cultural Factors in Women’s Labor Force Participation in Chile. Feminist Economics, 16(2):27-46.
Cuberes, D., and Teignier, M. (2014). Gender inequality and economic growth: A critical review. Journal of International Development, 26: 260-276.
Cunningham, M. (2008). Influences of Gender Ideology and Housework Allocation on Women’s Employment over the Life Course. Social Science Research, 37(1): 254-267. Duflo, E. (2012). Women’s empowerment and economic development. CEPR discussion
paper No. 8734. London: Centre for Economic Policy Research.
Fakih, A., and Ghazalian, P.L. (2015). Female employment in MENA’s manufacturing sector: the implications of firm-related and national factors. Economic Change and
Restructuring, 48: 37-69.
Farnsworth, B.B. (1976). Bolshevism, the Women Question, and Aleksandra Kollontai. The American Historical Review, 81(2): 292-316.
Fortin, N.M. (2005). Gender Role Attitudes and the Labour-market Outcomes of Women Across OECD Countries. Oxford Review of Economic Policy, 21(3): 416-438.
Gaddis, I., and Pieterse, J. (2012). Trade Liberalization and Female Labor Force Participation. Evidence from Brazil. IZA Discussion Paper No. 6809. Bonn: Institute for the Study of Labor.
Gaddis, I., and Klasen, S. (2014). Economic development, structural change, and women’s labor force participation: A reexamination of the feminization U hypothesis. Journal of Population Economics, 27: 639-681.
Gustafson, S.S., Wetzels, C.M.M.P., Vlasblom, J.D., and Dex, S. (1996). Women’s labor force transitions in connection with childbirth: A panel data comparison between Germany, Sweden and Great Britain. Population Economics, 9: 223-246.
Goldin, C. (1995). The U-shaped female labor force function in economic development and economic history. In Schultz, T.P. (Ed). Investment in Women’s Human Capital. Chicago, IL: Chicago University Press.
Human Development Reports. (2019). Retrieved from
http://hdr.undp.org/en/content/expected-years-schooling-females-years.
International Labor Organization. (2019). Retrieved from https://ilo.org/. Jaumotte, F. (2003). Female Labour Force Participation. Paris: OECD.
25 Khitarishvili, T. (2016). Gender Dimensions of Inequality in the Countries of Central Asia,
South Caucasus, and Western CIS. Working Paper No. 858. Annandale-on-Hudson, NY: Levy Economics Institute.
Klasen, S. (2002). Low schooling for girls, slower growth for all? Cross-country analysis on the effect of gender inequality in education on economic development. World Bank Economic Review, 16(3): 91-132.
Klasen, S. (2018). The Impact of Gender Inequality on Economic Performance in Developing Countries. Annual Review of Resource Economics, 10: 279-298.
Klasen, S., and Pieters, J. (2012). Push or Pull? Drivers of Female Labor Force Participation during India’s Economic Boom. IZA Discussion Paper No. 6395. Bonn: Institute for the Study of Labor.
Klasen, S., and Lamanna F. (2009) The impact of gender inequality in education and employment on economic growth: new evidence from a panel of countries. Feminist Economics, 15(3): 91-132.
Luci, A. (2009). Female labour market participation and economic growth. International Journal of Innovation and Sustainable Development, 4(2/3): 97-108.
Nam, S. (1991). Determinants of Female Labor Force Participation: A Study of Seoul, South Korea, 1970-1980. Sociological Forum, 6(4): 641-659.
Ngai, L.R., and Petrongolo, B. (2013). Gender gaps and the rise of the service economy. CEP discussion Paper 1204. London: Centre for Economic Performance.
Mammen, K., and Paxson, C. (2002). Women’s Work and Economic Development. Journal of Economic Perspectives, 14(4): 141-164.
Mitra, A. (2005). Women in the urban informal sector: perpetuation of meagre earnings. Development and Change, 36(2): 291-316.
Ofer, G., and Vinokur, A. (1983). The Labor-Force Participation of Married Women in the Soviet Union: A Household Cross-section Analysis. Journal of Comparative
Economics, 7: 158-176.
Olmested, J. (2011). Gender and Globalization: The Iranian Experience. In Bahramitash, R., and Esfehani, H.S. (Eds.),Veiled Employment: Islamism and the Political Economy of Women’s Employment in Iran: 25-52. Syracuse, NY: Syracuse University Press.
Rees, R., and Riezman, R.G. (2012). Globalization, gender and growth. Review of Income and Wealth, 58(1): 107-117.
Sackey, H.A. (2005). Female Labour Force Participation in Ghana, the Effects of Education. Nairobi: The Regal Press.
Sen, A.K. (1998). Development as capability expansion. Journal of Urban Planning and Development, 19: 41-58.
26 Tzannatos, Z. (1999). Women and Labor Market Changes in the Global Economy: Growth
Helps, Inequalities Hurt and Public Policy Matters. Social Protection Discussion Paper Series No. 9808. Washington, D.C.: The World Bank.
Usha, K.B. (2005). Political Empowerment of Women in Soviet Union and Russia: Ideology and Implementation. International Studies, 42(2): 141-165.
De Vries, G., Timmer, M., and de Vries, K. (2015). Structural transformation in Africa: Static Gains, Dynamic Losses. The Journal of Development Studies, 51(6): 674-688.
Woetzel, J., Madgavkar, A., Ellingrud, K., Labaye, E., Devillard, S., Kutcher, E., Manyika, J., Dobbs, R., and Krishnan, M. (2015). How Advancing Women’s Equality Can Add $12 Trillion to Global Growth. New York, NY: McKinsey Global Institute Report.
World Bank. (2019). Retrieved from https://data.worldbank.org/.
27 Appendix A
Table 9 : Correlation matrix
Variable FLPR Soviet dummy lnGDP per capita lnGDP per capita2 Fertility Size of the service industry Educational
attainment Government expenditure Trade openness
FLPR 1
Soviet dummy 0.0659 1
lnGDP per capita -0.2522 -0.0860 1
lnGDP per capita2 -0.2238 -0.1003 0.9975 1
Fertility 0.2059 -0.2384 -0.7684 -0.7476 1 Size of the service
28 Appendix B
Table 10: Robustness check
Dependent variable: female labor force participation rate Variable 1991-2018
(augmented) (augmented) 1991-1997 (augmented) 1998-2004 (augmented) 2005-2011 (augmented) 2012-2018 Soviet dummy 3.5517 (0.6687)*** (1.7110)* 2.9261 (1.4377)* 2.9745 (1.0977)*** 4.4312 (1.2400)*** 4.5871 lnGDP per capita -53.8899 (2.1610)*** (4.0627)*** -60.0167 (4.4421)*** -51.7382 (4.2013)*** -49.6581 (4.7514)*** -51.1247 lnGDP per capita2 (0.1150)*** 3.0562 (0.2186)*** 3.4605 (0.2324)*** 2.9599 (0.2238)*** 2.8340 (0.2531)*** 2.8765 Fertility -0.2339 (0.2976) (0.5212)* -0.9234 (0.5909) -0.4329 (0.6120) 0.8248 (0.7170) 1.0583 Size of the service industry (0.0266)*** -0.2596 (0.0555)*** -0.4120 (0.0533)*** -0.3054 (0.0470)*** -0.2039 (0.0572) -0.0372 Educational attainment (0.1161)*** 1.4347 (0.2238)*** 1.5532 (0.2365)*** 1.3150 (0.2428)*** 1.3862 (0.2724)*** 0.8421 Government expenditure (0.0471)*** -0.1645 (0.0804) 0.0893 (0.0961) -0.1678 (0.0977)*** -0.3684 (0.1091)*** -0.2987 Trade openness 0.0101 (0.0039)*** (0.0093)* 0.0165 (0.0090)* 0.0066 (0.0068) 0.0029 (0.0073) 0.0075 Constant 278.6379 (10.5392)*** (18.8499)*** 303.2049 (21.4698)*** 271.7696 (20.7755)*** 258.0973 (23.7039)*** 266.5354 Observations 3,786 925 1,201 1,067 773 ρ-value 0.0000 0.0000 0.0000 0.0000 0.0000 R-squared 0.3035 0.3829 0.3218 0.2858 0.2488
