The influence of the share of females in parliament on subjective and objective wellbeing

The influence of the share of females in parliament on

subjective and objective wellbeing

University of Groningen

Faculty of Economics and Business

MSc Economic Development & Globalisation

Master Thesis

Abstract: Female political engagement has lately earned increasing importance. It is said that many women in political power positions are able to lead countries more effective, more sustainable and more fruitful. In this study this assumption is tested by assessing the impact of increased female representation on two measures of wellbeing: Objective and subjective wellbeing. By doing this, this study examines whether female political engagement not only factors economic growth but also life satisfaction by the population. On a global level no significant results were able to be obtained. After splitting the country data into different geographical regions, most still experienced no significant impact, however one region experienced a negative effect on GDP per capita and life satisfaction while another region experienced a positive effect on GDP per capita. Thus, further research is needed in order to assess the consequences of increased female political participation on wellbeing.

Name Student: Inés Faghihi Amami

Student ID Number: S3242161

Student E-Mail: i.faghihi@student.rug.nl

Date Thesis: January 5, 2020

Table of Content

1 – Introduction ... 3

2 – Literature Review ... 5

2.1 – GDP as an objective measure of welfare ... 5

2.2 – Subjective wellbeing as a measure of welfare ... 5

2.3 – How female political engagement stimulates wellbeing ... 6

3 – Data and Methodology ... 8

3.1 – Model Specification ... 8

3.2 – The dependent variables ... 9

3.3 – The independent variable ... 10

3.4 – The control variables ... 10

3.5 – Descriptive Statistics and Correlation Matrix ... 11

4 – Empirical Results ... 13

5 – Robustness ... 18

6 – Conclusion and Limitations ... 22

7 – References ... 24

1 – Introduction

Already in 1999 Noble Prize winner Amartya Sen claimed in his book (p. 203) “Development as Freedom” that “[n]othing, arguably, is as important today in the political economy of development as an adequate recognition of political, economic and social participation and leadership of women. This is indeed a crucial aspect of ‘development as freedom.” Also recently, with global crises such as the Climate Crisis and the COVID-19 pandemic, the importance of the influence of women in politics has risen. Germany’s Chancellor Angela Merkel has been one of the greatest examples of how women can lead countries intro prosperity and lead as a world power. Also other nations’ heads of government, such as New Zealand’s prime minister Jacinda Ardern, have been examples of a fresh change of ruling that has brought their countries prosperity and growth.

Due to rising societal pressure and legislations such as quotas, there has been a significant global increase in the proportion of women in leadership positions and parliaments. Nevertheless, yet only around a quarter of all available seats in parliaments are taken by females (World Bank, 2020; Bhalotra, 2018). This underrepresentation of females in politics is just one indicator of how gender inequality is still a worldwide problem. Aside from the mere imbalance of gender representation, female political leadership specifically has gained more and more attention in recent years.

Much evidence has been collected on the influence of female political leaders on economic growth already. For example, Bhalotra (2018) found that female legislators in India were able to raise economic growth by 1.8 percentage points more than male legislators. Also Xu (2015) found that in recent years female political participation fostered economic growth in Asia.

Important to note is that these studies almost exclusively look at how female political engagement influences either GDP or GDP growth. GDP in this case is the most common variable used in order to establish a country’s welfare level. Rarely any research so far has gotten into the effects of female participation in politics on subjective wellbeing measures, such as happiness or life satisfaction. The reason why research should not only focus on objective wellbeing measures such as GDP growth is because it may only tell half of the story. “Increased female political engagement has led to GDP growth and thus more females should be in politics” is not a fully derived argument. The reason for that is due to the monetary biasness of the objective welfare measure GDP. Just because a country’s growth rate is high, does not necessarily mean the country’s overall welfare level is good. Many countries in the world are experiencing high levels of GDP or GDP growth, nevertheless often their population is experiencing high levels of income inequality and poverty.

This comparison of changes between subjective and objective wellbeing measures due to a stimulus, in this case the increase of females in parliament, shows a research gap that will be filled with this study.

Therefore, the research question of this paper will be: Does increased female representation in politics measured by the seats held by women in national parliaments increase objective and subjective welfare measures alike?

Using two fixed effects models and data from 157 countries over 14 years this paper empirically tests whether increased numbers of seats in parliament held by women have a positive impact on GDP per capita and life satisfaction as two different ways of measuring wellbeing.

With this in mind, the aim of this paper is to contribute to the existing literature on the relationship between female representation and wellbeing. Additionally, this thesis is aiming to add to existing empirical literature in two ways: Firstly it sheds light on a research area that has, to the best of this author’s knowledge, hardly been explored yet, namely the relationship between increased female government representation and subjective wellbeing as another measure of welfare. Secondly, this paper compares the influence of a stimulus on objective and subjective wellbeing in one analysis.

The results show that on a global level there is no significant effect of increased women in parliament on neither GDP per capita, as a measure of objective wellbeing, nor the Life Ladder Index for life satisfaction, as a measure of subjective wellbeing. When splitting the sample by geographical regions, still most regions experience no positive or negative effect by increased participation of females in politics. Nevertheless, there is a small positive significant influence of an increased share of women in parliament on GDP per capita in East Asia and the Pacific. On the other hand, however, the results also show a small negative significant influence of increased women in parliament on GDP per capita and the Life Ladder Index in Europe and Central Asia.

2 – Literature Review

2.1 – GDP as an objective measure of welfare

GDP is the most common indicator for assessing the welfare level of a country. There are several arguments in favour of using GDP as an indicator. It is standardised, monetarily accountable, and practical. It seem conclusive enough in order to argue that GDP is an appropriate measure of assessing welfare. The richer a country the better it is off. If one blindly accepts GDP as an objective measure of welfare this conclusion might be accurate, however no indicator is free of shortcomings and over the past many scholars have raised the issue that GDP seems to depict an incomplete picture of welfare and that subjective wellbeing measures are as important as objective ones (Nikolova, 2016). The reasons for that are multifold. GDP, the measure of the value of all goods and services provided in a nation, does reflect the economic activity of a nation, however does not capture whether this economic activity also mirrors the population’s wellbeing, i.e. the positive and negative emotions over a period of time (Nikolova, 2016). While technically it seems logical to connect increasing GDP levels with increasing levels of happiness and wellbeing, in practice this is not always the case. One example of that shows the “paradox of unhappy growth” by Graham and Lora (2009) that argues that rapid growth levels often bring insecurity and instability to people and thus deteriorating their subjective wellbeing (Graham and Lora, 2009; Nikolova, 2016). Another example is the famous “Easterlin paradox” that predicts that as countries get richer, they do not get subjectively happier. Additionally, economic “bads” such as natural catastrophes, pollution and wars account positively to GDP, while obviously having a negative impact on the population’s wellbeing, while for example volunteer work contributing to civic wellbeing is not caught by GDP (Nikolova, 2016). This lack of wellbeing measure also makes country comparisons difficult: Two countries with similar GDP levels may use their resources differently, so that in one nation only a small elite’s wellbeing is supported while in the other nation resources are used in order to promote the general population’s wellbeing (Nikolova, 2016).

Thus it becomes clear that one cannot simply draw all welfare conclusions based on GDP and when one is trying to assess the impact of a certain stimulus on the economy, another measure is needed to get a complete picture. This thesis complements the existing empirical evidence and literature by also including subjective wellbeing as a measure of welfare.

2.2 – Subjective wellbeing as a measure of welfare

There are several advantages coming alongside subjective wellbeing measures. First and foremost, different to GDP that only focuses on the income dimension, subjective wellbeing takes monetary and non-monetary dimensions into account which helps to get a clearer picture about the sources of certain levels of wellbeing (Nikolova, 2016).

2.3 – How female political engagement stimulates wellbeing

This study investigates the relationship between female representation in politics and the economic consequences from that. In specific, it tests whether the characteristics of higher shares of seats in parliament taken by women will have positive effects on welfare. This is a valid point of considerations since many studies have pointed out the fundamental differences in policy making between men and women. One needs to understand the channels through which female political representation indirectly may affect wellbeing.

Firstly, corruption levels tend to differ with females in political power positions. Research has shown that women in office tend to be less corrupt, less selfish and have higher levels of morality (Dollar et al., 2001; Swamy et al., 2001, Frank et al., 2011; Rivas, 2012; Paweenawat, 2008), which immediately translate into higher levels of economic welfare due to increased investment, reduced income inequality and lower levels of political instability (Mo, 2001; Dollar et al., 2001).

Secondly, women in political power are known for focussing mainly on social matters compared to males in political power. Virginia Tech (2018) for example shows that increased female political participation is associated with overall improvements in a nation’s health and education. One explanation for female politicians to focus on matters such as healthcare, education and welfare is given by Chaney (2014), claiming that women in political power positions regard their functions in parliament as an extension of their traditional roles as wives and mothers. This will consequently lead to higher levels of government spending social matters such as health care programs and education. Following, basic economic theory suggests that increased levels of government spending in those areas positively affects GDP and are essential to reduce income inequality which, consequently, results in higher economic welfare (World Inequality Report, 2018). Gender equality and the fight for equal rights between men and women also comes into play here. Female politicians are much more likely to ensure that females enjoy the same rights as males, such as equal access to education and the labour market. Especially giving females access to equal education has shown to have high returns on investment and significant potential for economic growth due to increased skilled human capital (Dollar and Gatti, 1999). Additionally, gender equality will increase and strengthen the overall labour force and thus foster economic growth (Blackden et al., 2007).

Since corruption is mainly causing political instability (Mo, 2001) which causes individuals to feel insecure, one can expect that lower corruption levels will lead to lower political instability and thus to higher levels of subjective wellbeing. This intuition is supported by research by Kabene et al. (2017) that empirically shows that the lower corruption levels associated with females in parliament result in higher levels of happiness. Additionally, the channel of gender equality can be expected to influence the subjective wellbeing, since more females will feel represented and heard which translates in higher levels of life satisfaction and confidence into the government. Lastly, social policies that are frequently observable with increased female political leadership can be expected to rise overall satisfaction and happiness. Especially the areas of education and health care will have significant impact on the overall wellbeing of a country.

Based on this research the hypotheses for this paper can now be set up.

Hypothesis 1: Higher levels of female representation in politics results in higher levels of objective welfare measured by GDP.

Hypothesis 2: Higher levels of female representation in politics results in higher levels of subjective welfare measured by life satisfaction perceived by the society.

3 – Data and Methodology

3.1 – Model Specification

This paper assesses the effect of one key independent variable, namely the extent of presence of women in parliament, on two different indicators of wellbeing, namely GDP per capita and life satisfaction. In order to do this, one baseline regression is computed which will be used in two different models. These two models only differentiate in the key dependent variables GDP per capita as a measure of objective wellbeing and the Life Ladder Indicator by the World Happiness Report as a measure of subjective wellbeing while keeping the baseline regression the same. With keeping the baseline regression equal for both models, this thesis will be able to assess whether forces, such as gender equality, pull equally on different measures of wellbeing.

The dataset includes a range of 157 countries all over the globe between the years 2005 and 2018. Appendix A – Country List shows an overview of all countries in different geographic regions categorised by the World Bank. These extensive panel data are allowing for examining cross-sectional and time-series data, using between and within country differences. The sample of countries and variables was restricted by the availability of the life ladder indicator by the World Happiness Report data, GDP per capita, and of the proportion of seats held by women in parliament. The dataset is not balanced, which means that some observations are not included in all models.

In the following, the two baseline models used for testing the hypotheses will be introduced.

Model 1: Objective Wellbeing

(1) !"#$_!" = '_# + '_$)"*+,_!"%&+ '_'-_!"+ ._! + /_"+ 0_!"

The above equation (1) corresponds to the model assessing the effect of the proportion of seats held by women in parliament on the objective wellbeing indicator GDP per capita, whereas !"#$ depicts the logged GDP per capita in current international dollars converted by purchasing power parity in country c at time t, '_# the slope of the regression intercept, '_$ the slope of the key independent variable )"*+, showing the percentage of seats held by women in parliament in country c at time t lagged by one year, '' the coefficient vector C, a summary

vector for all control variables, ._! the time-invariant country fixed effects capturing cultural, geographical and institutional differences, /_" the time-variant year fixed effects and 0 the error term.

Model 2: Subjective Wellbeing

(2) 1!" = '#+ '$)"*+,!"%&+ ''-!"+ .! + /"+ 0!"

depicts the Life Ladder index by the World Happiness report in country c at time t, '_# the slope of the regression intercept, '_$ the slope of the key independent variable )"*+, showing the percentage of seats held by women in parliament in country c at time t lagged by one year, '_' the slope of vector -, a summary vector for all control variables, ._! the time-invariant country fixed effects capturing cultural, geographical and institutional differences, /_" the time-variant year fixed effects and 0 the error term.

For comparison purposes these models do not differ apart from the dependent variable.

Regarding the estimation of these models there are important issues arising that need to be addressed. Firstly, the question arises why a fixed effects model was used instead of a random effects model. In order to answer this question the Hausman test was used for both models. Each time the Hausman test reports a prob>chi2 value of 0.000. Thus, the hypothesis of no systematic differences in the coefficients will be rejected and country and year fixed effects are favoured over random effects. Using fixed effects helps in the second issue that needs to be addressed: endogeneity. A fixed effects model with year and country fixed effects helps in tackling the issue of endogeneity as it may be controlling for omitted variables. In regression equations such as those used in this paper with dependent variables such as GDP and life satisfaction it is impossible to control for all relevant variables. Additionally, the more control variables are included the smaller the number of observations get. In order to not reduce the already significantly shrank dataset, the baseline regressions only include a few selected control variables. In the robustness check section however, more control variables will be used for the two models.

Thirdly, the issue of lagging the variable women needs to be discussed. As it will be explained below, the variable women is lagged by one year in order to capture the time needed to observe societal changes due to changes in the government. Nevertheless, a one year lag is rather arbitrarily chosen, especially since the process of passing bills and implementing changes can be varying between countries. Thus, the robustness checks include the same estimation however with lagging the variable women by two, three, four, and five years.

Lastly, the issue of heteroskedasticity arises. Both models are tested for heteroskedasticity with the modified Wald Test. The tests for both models show a prob>chi2 value of 0.000. In both cases this shows the presence of heteroskedasticity which will be corrected with robust standard errors.

3.2 – The dependent variables

Hypothesis one predicts that an increase of females in parliament increases GDP per capita. If this hypothesis turns out to be proven correct, model (1) will predict an increase in the logged GDP per capita when the variable women increases.

The second dependent variable of interest for model specification (2) is the Life Ladder Index by the World Happiness Report. The Life Ladder Index is a measure of life satisfaction evaluation which asks survey respondents to place their live on a “ladder” ranging from 0 to 10. 0 depicts the worst and 10 the best possible life (Helliwell, Layard & Sachs, 2019). The average score of each respondent make up the Life Ladder index per year per country. Thus, the higher the value, the more satisfied the average population of a certain country was in a specific year. This measure of wellbeing is very different from GDP per capita as it is fully based on self-reported data and is not measured by internationally standardised characteristics. Hypothesis 2 predicts that an increase of women in parliament also increases overall life satisfaction. Thus, if the variable women increases, the Life Ladder variable should also increase if this hypothesis is proven to be correct.

3.3 – The independent variable

The key observation to be made for this analysis is the impact of the share of females in parliament on wellbeing. For that, the key independent variable needed is the percentage share of seats in parliament held by women. The data is obtained by the Inter-Parliamentary Union (IPU) which measures the proportion of seats held by women in national parliaments in percent. Important to note is that the key independent variable is lagged by one year. The reason for that is that the process of electing and installing government representatives takes time, and that these government representatives also need time to implement changes that again need time to show results that can be perceived by the population. For example: If in country X there is an increase from 15% of females holding seats in parliament to 30% in year 0, the effects of this 15% increase are likely to only show in year 1. That is due to bills and proposals first having to be made and implemented. The initial effect of a passed bill on increased education expenditure will most likely only be small and only translate into e.g. happiness of future improved learning. The actual effect however will most likely only be experienced by the ones that receive the better quality education after the implementation of the bill. Even if there is an initial effect of satisfaction or dissatisfaction by the population only based on the gender representation changes, the lag of the variable will either capture an unchanged (dis)satisfaction or a more accurate, adapted vision of the parliament status.

3.4 – The control variables

The first control variable used in the baseline regression is the total unemployment level as a percentage of the total labor force modelled by the International Labor Organisation. The reasoning behind is that increased unemployment levels both influence GDP per capita and life satisfaction. By Okun’s law it is established that an increase in unemployment a decrease in GDP can be expected due to the lower levels of output generated by employed labor. Similarly, it is to assume that higher levels in unemployment result in lower levels of life satisfaction due to the lack of compensation, social contact and feeling of usefulness (Theodossiou, 1998).

Secondly, human capital is an important factor that needs to be controlled for. Increased human capital, i.e. knowledge, health and skills of the labor force, increases productivity and thus GDP. Additionally, better health and education also increases overall life satisfaction. Therefore, two more control variables are included: Life expectancy at birth as a proxy for health and gross enrollment rates for secondary education as a proxy for education.

Life expectancy at birth in years is indicating the number of years a newborn infant would live if prevailing patterns of mortality at the time of its birth were to stay the same throughout its life. The gross enrollment rates in secondary education is the ratio of the total enrollment regardless of age to the population of the age group that officially corresponds to the level of education shown (World Bank, 2020). Secondary education completes the basic education aiming for laying the foundations for lifelong learning and human development (World Bank, 2020). Both of these indicators are retrieved from the World Bank.

Lastly, the real interest rate is included as a control variable by the World Bank. By standard economic theory interest rates affect economic development through e.g. the cost of borrowing, disposable income, and the value of currencies. However, this also influences life satisfaction through the channels of disposable income and of mortgage borrowing. Rohe & Stegman (1994) found in their study, that home owners experienced significant higher levels of life satisfaction. Since in most cases mortgages are needed in order to be able to become a home owner, interest rates have a direct effect on the ability to afford a mortgage and thus the ability to be a home owner.

3.5 – Descriptive Statistics and Correlation Matrix

As a preliminary step of the estimation the following table presents the descriptive statistics of the data.

Table 1: Descriptive Statsistics

Variable Obs Mean Std. Dev. Min Max

Log GDP/Capita 1,585 9.294021 1.16203 6.404971 11.86101 Life Ladder Index 1,614 5.457628 1.130993 2.661718 8.018934

Female share 1,614 20.5502 11.11956 0 63.75

Interest rate % 1,087 6.176554 9.001475 -34.46203 61.8826

Gross enrollment 895 87.10835 29.25949 0 163.9347

The data shows that for most variables the observations are balanced. However, the gross enrolment rate for secondary education shows the lowest observations with splitting the dataset in almost half.

The country with the lowest log GDP per capita is the Democratic Republic of Congo in 2009 with 6.40 and the country with the highest log GDP per capita is Qatar in 2012 with 11.86. In terms of the Life Ladder Index, the country reporting the lowest value is Afghanistan in 2017 with 2.66 and the country reporting the highest value is Denmark in 2005 with 8,01. Both these variables show that there is a wide gap between countries in different wellbeing measures. When it comes to the share of seats held by females in parliament several countries such as Qatar, Saudi Arabia and Yemen report no females present at all. The highest share reported is 63.75 by Rwanda between 2013 and 2016.

Additionally, table 2 and 3 show the correlation matrixes for both models:

Table 2: Correlation Matrix Model 1

Variables (1) (2) (3) (4) (5) (6) (1) Log GDP/Capita 1.000 (2) Female share 0.046 1.000 (3) Unemployment 0.083 -0.001 1.000 (4) Life Expectancy 0.884 0.119 0.093 1.000 (5) Interest rate % -0.279 -0.037 -0.028 -0.151 1.000 (6) Gross enrollment 0.756 0.138 0.211 0.751 -0.308 1.000

Table 3: Correlation Matrix Model 2

Variables (1) (2) (3) (4) (5) (6) (1) Life Ladder Index 1.000

(2) Female share 0.176 1.000

(3) Unemployment -0.140 0.001 1.000

(4) Life Expectancy 0.707 0.116 0.087 1.000

(5) Interest rate % -0.219 -0.036 -0.032 -0.146 1.000

(6) Gross enrollment 0.543 0.134 0.204 0.753 -0.303 1.000

4 – Empirical Results

In order to get a first oversight of the potential relationship between GDP per capita and the share of female seats in parliament a scatterplot with fitted values were created.

Figure 1: Scatterplot log GDP per capita against the share of female seats in

parliament

Figure 2: Scatterplot Life Ladder Index against the share of female seats in

parliament

Figure 1 and 2 show a first indication of what can be expected from the estimation results. The correlation coefficient visually depicted above between the log of GDP per capita and the share of females in parliament is 0.1362. The correlation between the Life Ladder Index against share of females in parliament is 0.2760. Although both of these values are not very strong, it gives a first indication of a positive influence of the share of females in parliament on wellbeing.

To test these assumption empirically, the effect of the female share in parliament is tested against the two measures of wellbeing: Log GDP per capita and the Life Ladder Index. The results are depicted below in table 4.

Table 4: Share of women in parliament on Log GDP per capita and Life Ladder Index

(1) (2)

VARIABLES

Log GDP per

capita (PPP) Life Ladder Index

Share of women in parliament (n-1) 0.002 -0.006

(0.002) (0.007)

Unemployment Level -0.014*** -0.070***

(0.004) (0.017)

Life Expectancy at Birth -0.006 -0.055

(0.011) (0.057)

Gross School Enrollment (%) 0.001* 0.003** (0.000) (0.001) Constant 9.566*** 9.810** (0.808) (4.103) Observations 465 469 R-squared 0.784 0.153 Number of countries 88 88

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

Column (1) presents the results of model 1 in which the women indicator is tested against the log GDP per capita. Column (2) presents the results of model 2 in which the Life Ladder Index is the dependent variable.

Against the earlier predictions from the scatterplot and literature, it seems like that the share of women in parliament has no significant effect on neither the log of GDP per capita or the Life Ladder Index.

Additionally, most of the control variables do show a significant influence on the dependent variables very much in line with the theory. It can be seen that in both models higher unemployment levels will reduce wellbeing, and that higher enrolments in secondary education will increase wellbeing. A higher real interest rate also reduces the log GDP per capita as expected by economic theory. Life expectancy at birth as a proxy of health does not show a significant result.

In general, one might conclude from this that the amount of females in parliament has no significant effect on any wellbeing measure. However, one needs to bear in mind, that there are geographical, cultural, historical, political and legal differences in all these countries that may have a huge influence on the impact of females in parliament and balance each other out. Countries with a rather traditional view not representing many women in parliament may overshadow other countries due to the size of the dataset. Thus, in order to test this, the dataset was split in geographical regions as indicated by the World Bank: East Asia & Pacific, Europe & Central Asia, Latin America & Caribbean, Middle East & North Africa, North America, South Asia, Sub-Saharan Africa (see appendix for the country lists).

Table 5: Share of women in parliament on Log GDP per capita, geographic regions East Asia & Pacific Europe & Central Asia Latin America & Carribean

Middle East &

North Africa North America South Asia

Sub-Saharan Africa

VARIABLES GDP per capita (PPP)

Share of women in parliament (n-1) 0.005*** -0.007*** 0.002 0.011 0.004 0.003 -0.004

(0.002) (0.002) (0.001) (0.009) (0.000) (0.006) (0.006)

Unemployment Level -0.014** -0.021*** -0.015** 0.009 -0.030 -0.053** -0.002

(0.006) (0.004) (0.006) (0.016) (0.000) (0.018) (0.017)

Life Expectancy at Birth -0.017 -0.039* 0.028* 0.060 0.115 -0.105* -0.014

(0.013) (0.021) (0.016) (0.082) (0.000) (0.056) (0.013)

Real Interest Rate -0.001 -0.003** 0.000 -0.009*** -0.024 -0.003 -0.001

(0.002) (0.001) (0.001) (0.003) (0.000) (0.003) (0.003)

Gross School Enrollment (%) 0.001 -0.005*** 0.001** 0.001 -0.033 0.014*** 0.001

(0.001) (0.002) (0.000) (0.005) (0.000) (0.003) (0.002)

Constant 11.687*** 12.751*** 7.514*** 6.087 5.048 13.702*** 8.433***

(0.995) (1.610) (1.175) (6.112) (0.000) (3.344) (0.723)

Observations 56 132 131 36 14 29 67

R-squared 1.000 0.996 0.986 0.999 1.000 0.999 0.996

Table 6: Share of women in parliament on Life Ladder Index, geographic regions

East Asia & Pacific

Europe & Central Asia

Latin America & Carribean

Middle East &

North Africa North America South Asia

Sub-Saharan Africa

VARIABLES Life Ladder Index

Share of women in parliament (n-1) 0.019 -0.022** -0.005 0.086 -0.042 0.045 -0.005

(0.023) (0.010) (0.005) (0.057) (0.000) (0.076) (0.031)

Unemployment Level -0.040 -0.076*** -0.084*** -0.062 -0.024 0.192 -0.035

(0.086) (0.016) (0.027) (0.100) (0.000) (0.211) (0.096)

Life Expectancy at Birth -0.044 -0.102 0.191*** -0.707 0.365 -0.823 -0.196**

(0.183) (0.087) (0.065) (0.517) (0.000) (0.657) (0.072)

Real Interest Rate -0.017 -0.010* -0.001 0.015 0.130 -0.005 -0.023

(0.022) (0.005) (0.006) (0.016) (0.000) (0.033) (0.018)

Gross School Enrollment (%) 0.015* -0.004 0.003 0.027 0.143 -0.034 0.001

(0.008) (0.007) (0.002) (0.030) (0.000) (0.038) (0.010)

Constant 8.552 13.905** -7.507 56.095 -36.769 52.047 14.674***

(14.180) (6.817) (4.859) (38.734) (0.000) (39.342) (3.980)

Observations 56 132 134 36 14 29 68

R-squared 0.930 0.950 0.883 0.960 1.000 0.911 0.851

With the geographical differences there are significant differences to be seen in the two models. For model 1 (table 5) it can be seen that column 1 and 2 report highly significant results for the influence of the share of females in parliament on the log GDP per capita. Column 1 reports that a 1% increase of females in parliament increases the log GDP per capita by 0.005% in the East Asian and Pacific region. Contrarily, column 2 reports that an increase by 1% of females in parliament decreases the log GDP per capita by 0.007% in Europe and Central Asia. In all other regions in columns 3-7 there are yet again no significant impacts to be denoted.

The second model (table 6) shows similar results. Column 2 presents that a 1% increase of women in parliament reduces the Life Ladder Index by 0.022 units. In the meantime, columns 1 and 3-7 report no significant impact of the share of women on the Life Ladder Index.

These results have a few implications. Firstly, even after splitting the countries in geographical regions, in most regions there is no effect of women in parliament on either subjective or objective wellbeing. In the East Asia & Pacific regions, an increase of women in parliament has a positive and significant effect on log GDP per capita. It does not, however, show that increased female participation in politics also positively influence subjective wellbeing. In Europe and Central Asia there is an effect of women in parliament on both objective and subjective wellbeing. However, this effect is negative.

The reason why especially the latter result was obtained may be due to the split of the geographic regions. The European and Central Asian group is the biggest country group of the seven. Especially Europe has many different cultures and political environments. Similar problems as with the global analysis may be the reason why these contradicting results have shown.

5 – Robustness

As mentioned earlier, there will be two extensions for robustness checks to the models. Firstly, the two models will be expanded by more control variables, and secondly, there will be more lags of the variable women introduced.

1. Including more control variables

Different than to the baseline models, this section will include more control variables to each of the baseline models with the difference that they do not have to be shared with both of them. This means, there will be a new control variables for model 1 and new control variables for model 2. That will lead to the fact that those two regressions will not be comparable, however for the robustness check this is not of importance.

For model 1, one new control variables will be added. This is gross savings as percentage of GDP, defined as gross national income less total consumption, by the World Bank will be added.

For model 2, GDP per capita will be included, as the prosperity of the country one is living in can have an impact on overall life satisfaction.

The results can be found in table 7.

Table 7: Share of women in parliament on Log GDP per capita and Life Ladder Index

(1) (2)

VARIABLES Log GDP per Capita Life Ladder Index

Share of women in parliament (n-1) 0.002 -0.007

(0.002) (0.007)

Unemployment Level -0.012** -0.067***

(0.005) (0.016)

Life Expectancy at Birth -0.006 -0.069

(0.011) (0.054)

Real Interest Rate -0.004*** -0.005

(0.001) (0.004)

Gross School Enrollment (%) 0.001* 0.003***

(0.000) (0.001)

Gross Savings (% of GDP) 0.003

(0.002)

log GDP per Capita -0.000

(0.000)

Constant 9.410*** 10.538***

(0.752) (3.871)

R-squared 0.790 0.167

Number of country 87 87

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

Even with including more control variables to avoid the omitted variable bias, the results barely differ from the previous base line regressions.

2. Lagging the share of females in parliament

As explained earlier, the effect of political change is almost never immediate. Since different legislations and political environments also yield different times until change can be implemented, choosing any number is arbitrary. In order to assess whether the results stay equal or similar even with further lagging the key independent variable, table 8 shows the results for equation 1 with lagging the variable women 1, 2, 3, 4, and 5 years.

Table 8: Share of women in parliament on Log GDP per capita. 5 year lags

(1) (2) (3) (4) (5)

VARIABLES Log GDP per Capita

Share of women in parliament

(n-1) 0.002

(0.002) Share of women in parliament

(n-2) 0.000

(0.002) Share of women in parliament

(n-3) 0.000

(0.003) Share of women in parliament

(n-4) -0.000

(0.003) Share of women in parliament

(n-5) -0.001

(0.004)

Unemployment Level -0.014*** -0.012*** -0.014*** -0.010* -0.006

(0.004) (0.005) (0.005) (0.005) (0.005)

Life Expectancy at Birth -0.006 -0.016 -0.012 -0.009 -0.020

(0.011) (0.012) (0.014) (0.015) (0.015)

Real Interest Rate -0.005** -0.005** -0.005*** -0.005** -0.005**

(0.002) (0.002) (0.002) (0.002) (0.002)

Gross School Enrollment (%) 0.001* 0.000 0.001** 0.001 0.000

(0.000) (0.000) (0.001) (0.000) (0.000)

Observations 465 432 397 355 316

R-squared 0.784 0.769 0.732 0.726 0.708

Number of country 87 85 82 82 81

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

As it can be seen, while some of the significance levels of the control variables have changed, the significance, or rather the insignificance, of the key independent variable has not changed. Thus it can be expected that the results of the initial baseline regression model 1 can be confirmed.

Table 9 shows the same principle as before but for the model specification 2.

Table 9: Share of women in parliament on Life Ladder Index. 5 year lags

(1) (2) (3) (4) (5)

VARIABLES Life Ladder Index

Share of women in parliament

(n-1) -0.006

(0.007) Share of women in parliament

(n-2) -0.007

(0.007) Share of women in parliament

(n-3) -0.003

(0.006) Share of women in parliament

(n-4) 0.006

(0.006) Share of women in parliament

(n-5) -0.001

(0.010)

Unemployment Level -0.070*** -0.049** -0.068*** -0.070*** -0.083***

(0.017) (0.019) (0.021) (0.022) (0.022)

Life Expectancy at Birth -0.055 -0.065 -0.072 0.064 0.059

(0.057) (0.058) (0.073) (0.064) (0.075)

Real Interest Rate -0.002 -0.005 0.001 -0.001 -0.003

(0.003) (0.003) (0.004) (0.003) (0.003)

Gross School Enrollment (%) 0.003** 0.002** 0.001 0.002** 0.003**

(0.001) (0.001) (0.002) (0.001) (0.001)

Constant 9.810** 10.429** 11.063** 1.298 1.922

(4.103) (4.175) (5.258) (4.566) (5.374)

R-squared 0.153 0.130 0.130 0.127 0.163

Number of country 88 85 82 82 81

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

Again, it can be seen that while some of the significance levels of the control variables have changed, the coefficient of the share of women in parliament on the Life Ladder Index has no significant impact.

6 – Conclusion and Limitations

Due to the recent interest in women in leadership and power positions and their impact on a country’s wellbeing status more and more focus has shifted on gender equality in power positions. However, not much is yet known about the differences of female political leadership vs. male political leadership and the long term outcomes on wellbeing levels due to different leadership styles. Additionally, an old discussion comes into play, questioning the right way to assess a country’s wellbeing level. What makes a country “well?”. GDP has been a common measure of wellbeing for a long time. Nevertheless, it is also a known fact that it is not free of shortcomings. High levels of GDP do not necessarily mean that everyone is equally of. Enormous income gaps, environmental pollution, or misfunctioning institutions are some of the examples that may not be reflected in GDP, but affect the wellbeing of citizen. Here is where the measure of subjective wellbeing, i.e. the positive and negative emotions over time, becomes important.

In this study it is assessed whether an increase of females holding parliamentary seats has a positive influence on a country’s wellbeing level. In order to reduce the biasness towards one measure of wellbeing, this study combines subjective and objective wellbeing measures in a comparative study.

This paper compares 157 countries between 2005-2018. It fills a research gap by assessing the influence of women in parliament on two different measures of wellbeing. The results show that on a global level there is no significant effect of increased women in parliament on neither GDP per capita, as a measure of objective wellbeing, nor the Life Ladder Index for life satisfaction, as a measure of subjective wellbeing. When splitting the sample by geographical regions, still most regions experience no positive or negative effect by increased participation of females in politics. Nevertheless, there is a small positive significant influence of an increased share of women in parliament on GDP per capita in East Asia and the Pacific. On the other hand, however, the results also show a small negative significant influence of increased women in parliament on GDP per capita and the Life Ladder Index in Europe and Central Asia. Because of these results, both hypotheses cannot fully be accepted or rejected. The majority of the data however shows that on a global level, increased female seats in parliament do not have an impact on a country’s objective or subjective wellbeing level.

Even though this analysis goes in depth with a wide range of data, variables and makes contributions to existing literature, there are still some limitation that need to be taken into consideration.

should focus on finding trends on smaller regional subsamples in order to receive more concrete results.

7 – References

Alvaredo, F., Chancel, L., Piketty, T., Saez, E., & Zucman, G. (Eds.). (2018). World inequality report 2018. Belknap Press

Bhalotra, S. (2018). Are women politicians good for economic growth? - IGC. Retrieved 13 October 2020, from https://www.theigc.org/blog/are-women-politicians-good-for-economic-growth/

Binder, M. (2013). Subjective Wellbeing Capabilities: Bridging the Gap Between the Capability Approach and Subjective Wellbeing Research. Journal Of Happiness Studies, 15(5), 1197-1217. doi: 10.1007/s10902-013-9471-6

Blackden, M., Canagarajah, S., Klasen, S., Lawson, D. (2007). “Gender and Growth in Africa: Evidence and Issues,” in George Mavrotas and Anthony Shorrocks, eds. Advancing Development: Core Themes in Global Economics, pp. 349–70. London: Palgrave Macmillan. Dollar, D., Gatti, R. (1999). Gender inequality, income, and growth: are good times good for women? (Vol. 1). Washington, DC: Development Research Group, The World Bank.

Dollar, D., Fisman, R., & Gatti, R. (2001). Are women really the “fairer” sex? Corruption and women in government. Journal of Economic Behavior & Organization, 46(4), 423-429. Chaney, E. (2014). Supermadre: Women in politics in Latin America. Austin, Tex: Univ. of Texas Press.

Frank, B., Lambsdorff, J.G. and Boehm, F. (2011) Gender and corruption: lessons from laboratory corruption experiments. European Journal of Development Research 23(1), 59–71. Graham, C. (2009). Happiness Around the World: The Paradox of Happy Peasants and Miserable Millionaires. Oxford: Oxford University Press

Graham, C. and Lora, E. (2009). Paradox and Perception: Measuring Quality of Life in Latin America. Washington, D.C.: Brookings Institution Press.

Graham, C. (2011). Adaptation amidst Prosperity and Adversity: Insights from Happiness Studies from around the World. The World Bank Research Observer, 26(1), 105-137. Graham, C. (2011). Pursuit of Happiness: An Economy of Wellbeing. Washington DC: Brookings Institution Press.

Hellowell, J., Layard, R., & Sachs, J. (2019). World Happiness Report. Retrieved from https://s3.amazonaws.com/happiness-report/2019/WHR19.pdf

Kabene, S., Baadel, S., Jiwani, Z., & Lobo, V. (2017). Women in Political Positions and Countries’ Level of Happiness. Journal of International Women's Studies, 18(4), 209-217. Mo, P. (2001). Corruption and Economic Growth. Journal Of Comparative Economics, 29(1), 66-79. doi: 10.1006/jcec.2000.1703

Nikolova, M. (2016). IZA and the Brookings Institution. Retrieved from https://www.iza.org/publications/dp/10088

OECD. (2011). How's Life?. How's Life?. doi: 10.1787/9789264121164-en

Paweenawat, S. W. (2018). The gender‐ corruption nexus in Asia. Asian‐ Pacific Economic Literature, 32(1), 18-28.

Rivas, M. F. (2013). An experiment on corruption and gender. Bulletin of Economic Research, 65(1), 10-42

Rohe, W., & Stegman, M. (1994). The Effects of Homeownership: on the Self-Esteem, Perceived Control and Life Satisfaction of Low-Income People. Journal Of The American Planning Association, 60(2), 173-184.

Samuelson, Paul A., & Nordhaus, William D. (1995). Economics. New York: McGraw Hill. Stiglitz, J., Sen, A. and Fitoussi, J. (2009). The measurement of economic performance and social progress revisited. Reflections and overview. Commission on the Measurement of Economic Performance and Social Progress, Paris.

Swamy, A., Knack, S., Lee, Y., & Azfar, O. (2001). Gender and corruption. Journal of development economics, 64(1), 25-55

Tech, V. (2018). Study finds less corruption in countries where more women are in government. Retrieved 13 October 2020, from

https://www.sciencedaily.com/releases/2018/06/180615094850.htm

Theodossiou, I. (1998). The effects of low-pay and unemployment on psychological wellbeing: a logistic regression approach. Journal of Health Economics, 17, 85–104. World Bank. (2020). Proportion of seats held by women in national parliaments (%) | Data. Retrieved 13 October 2020, from https://data.worldbank.org/indicator/SG.GEN.PARL.ZS World Bank. (2020). School enrollment, secondary (% gross)) | Data. Retrieved 30 December 2020, from https://data.worldbank.org/indicator/SE.SEC.ENRR

Xu, L. (2015). Effects of Female Political Participation on Economic Growth: Evidence from Asian Countries. Retrieved 5 January 2021, from

8 – Appendix

Appendix A – Country List by Geographical Regions

East Asia & Pacific

Europe & Central Asia Latin America &

Caribbean

Middle East & North Africa

North America

South Asia Sub-Saharan Africa