Culture, Gender and Economic development

(1)

Msc. International Economics & Business

January, 2007

Culture, Gender and

Economic development

To what extent can the gender gap be explained by

cultural differences?

E.M. van den Berg

Dr. G. Peli

(2)

Abstract

The gender gap exists in every world country. However, the magnitude of the gender gap differences. This study follows up on research done by Augusto Lopez-Claros and Saadia Zahidi (2005), two economists, by World Economic Forum order. That research, called Women’s empowerment: measuring the global gender gap, constructed an index to measure the gender gap. This gender gap index is composed of five important dimensions; economic participation, economic opportunity, political empowerment, educational attainment, and health and well-being. My research investigates to what extent this gender gap can be explained by cultural differences and what the relation is between gender inequality and economic development. As a cultural measure, I use the four Hofstede dimensions; power distance index, individualism, masculinity and uncertainty avoidance. I performed multivariate linear regressions on a sample of 43 countries throughout the world. First, it turned out that three out of four Hofstede dimensions show a significant relationship to gender inequality. Power distance has a positive relation to gender inequality; individualism has a negative relation to gender inequality; and masculinity again a positive relation to gender inequality. Next, I investigated the relation between the gender gap and economic development. A country’s level of Gross National Income supposes to correlate with gender equality. However, the causality is not made clear. Unfortunately, the average percentage of GDP1 growth rate does not show positive causal relations with gender equality. Notwithstanding, in my opinion, reduction of the gender gap stays an important point of interest.

1

(5)

1. Introduction

1.1 Problem statement

In 2005, Saadia Zahidi and Augusto Lopez-Claros (2005), two economists at the World Economic Forum (WEF) did a study called ‘Women’s empowerment: Measuring the global gender gap’. This study was a first attempt to assess the current size of the gender gap by measuring the extent to which women in 58 countries have achieved equality with men. They address in this report five critical areas: economic participation, economic opportunity, political empowerment, educational attainment, and health and well-being.

The reason why this survey of the global gender gap is so important, is that unique data of annual Global Competitive Reports show that countries that do not fully take advantage of one half of the talent in their population are misallocating their human resources and thus undermining their competitive potential. The goal of the study is to create awareness by measuring the magnitude of the problem and to provide countries with a benchmarking tool. The authors hope that governments and NGOs can use the rankings of the gender gap index to identify issue areas and to learn from other countries experiences that have been successful in narrowing the gender gap.

The past three decades have observed a steadily increasing awareness of the need to empower women through measures to augment social, economic and political equity, to broaden access to fundamental human rights, and to improve nutrition, basic health and education. Gender refers to both women and men, and to their status, relative to each other. Gender equality refers to that stage of human social development when both men and women realize their full potential and opportunities, responsibilities and rights of individuals will not be determined by the fact of being born male or female.

The data show that in general the Nordic countries, Sweden, Norway, Denmark, Iceland, and Finland rank the highest among countries that provide women with economic freedoms and empowering them politically in addition to offering excellent healthcare and education.

(6)

research gives us insights into country’s cultures by making use of the five cultural dimensions; power distance, individualism, masculinity, uncertainty avoidance, and long-term orientation. Hofstede’s country sample consists of 80 countries. For a certain country, one’s culture can be indicated by numbers on the several dimension indexes. From now on I call this study the Hofstede study. In my study, I only use the first four dimensions. From the 58 countries selected by Zahidi and Lopez- Claros, 43 countries are overlapping with the 80 countries from the Hofstede study, which form the country sample for this study (see appendix 1,5, and 6). For the fifth dimension, long-term orientation, Hofstede gives only data for 33 out of the 80 countries. Only a few of those countries are overlapping with the gender gap index countries, so due to lack of data, I decided to leave the last dimension out of this research.

In my thesis, my intention is to find out to what extent this global gender gap can be explained by cultural differences?

Since the importance of narrowing the gender gap has already pointed out by several world economic committees, I would also like to address the causality of gender gap and economic development. In this second part of my study I pay attention to influence of the gender gap on a country’s economic development and on the relationship the other way around. What is the influence of a country’s economic development on the magnitude of the gender gap?

1.2 Research objectives

(7)

To what extent can the global gender gap be explained by cultural differences?

v To what extent can each of individual components of the gender gap index be

explained by cultural differences?

v Do economic growth and economic development have an influence on the magnitude

of the gender gap in a country?

v Has the gender gap an influence on the economic growth and development of a

country?

To answer them, I follow a two-step process. First, I address the relationship between culture and the gender gap. I will briefly describe the study by Professor Geert Hofstede, who developed a scale that can indicate a country’s score on each of four cultural dimensions. I will then extensively describe the Gender gap study by WEF2 order and how it is used related to Hofstede’s study. I will set up a number of hypotheses concerning the four cultural dimensions and the gender gap index. This analysis will be supported by a quantitative analysis of the effects of culture on gender inequality.

Second, I address the relationship between the gender gap and economic growth and economic development. I will find out and set up hypotheses concerning the level of GNI and percentage of GDP growth related to the gender gap index. Especially in this part, I pay attention to issues of causality. Does economic growth and development have an influence on the magnitude of the gender gap or is it the other way around?

After this introduction, chapter 2 contains the necessary background information on Geert Hofstede’s cultural dimensions and the current status of women in general. It also goes deeper into the partial subjects, describing women and the economy, women and politics, women and education and women and health. Those parts are described to help to interpret the outcome of this research.

Chapter 3 reviews the available literature on Geert Hofstede’s dimensions and the critique on his work. Moreover, it reviews literature on issues of women related to the above-mentioned partial subject as well as issues on economic growth and economic development.

Chapter 4 contains the hypotheses. Chapter 5 describes the methodology I use to answer my hypotheses.

2

(8)

Chapter 6 analyzes regression equations’ results for part one and part two of this research. First, six dependent variables are regressed on four explanatory ones. Next, the overall gender gap index (GGI) has investigated in both directions as described in the box above.

(9)

2. Background

2.1 Hofstede’s cultural dimensions

Professor Geert Hofstede conducted perhaps the most comprehensive study which gives us insights into other cultures so that we can be more effective when interacting with people in other countries. He is professor of organizational anthropology and international management at the University of Limburg at Maastricht in the Netherlands. He is also director of the Institute for Research on Intercultural Cooperation. Moreover, he has lectured and published internationally and acts as an international consultant to companies and organizations. In this section, I will briefly explain what he has basically done.

Geert Hofstede first ideas were based on a large research project into national culture differences across subsidiaries of IBM, a multinational corporation in 66 countries. Hofstede´s primary data were extracted from a pre-existing bank of employee attitude surveys undertaken around 1967 and 1973. He statistically analyzed the answers to these survey questions. That analysis revealed that there are four central and ‘largely dependent’ bi-polar dimensions of a national culture and that 40 out of the 66 countries in which the IBM subsidiaries were located could be given a comparative score on each of these four dimensions.

Subsequent studies covered students in 23 countries, elites in 19 countries, commercial airline pilots in 23 countries, up-market consumers in 15 countries, and civil service managers in 14 countries. The results of these studies were summarized in five independent dimensions of national culture differences, the four mentioned above, plus a fifth dimension long-term versus short-term orientation.

Hofstede defines these dimensions as follows.

Ø Power distance index (PDI): focuses on the degree of equality, or inequality, between people in the country’s society. A high power distance ranking indicates that inequalities of power and wealth have been allowed to grow within the society. A low power distance ranking indicates the society de-emphasizes the differences between citizen’s power and wealth. (Hypothesis 1 will utilize this dimension)

(10)

Individualism ranking indicated that individuality and individual rights are very important. A low individualism ranking typifies societies of a more collectivist nature with close ties between individuals. (Hypothesis 2 will utilize this dimension)

Ø Masculinity versus Femininity (MAS): focuses on the degree the society reinforces, or does not reinforce, the traditional masculine work role model of male achievement, control and power. A high masculinity ranking indicates the country experiences a high degree of gender differentiation. A low masculinity ranking indicates the country has a low level of differentiation and discrimination between genders. In these cultures, females are treated equally to males in all aspects of the society.

(Hypothesis 3, see chapter 4, will test this relationship)

Ø Uncertainty avoidance index (UAI): focuses on the level of tolerance for uncertainty and ambiguity within the society. A high uncertainty avoidance ranking indicates the country has a low tolerance for uncertainty and ambiguity. A low uncertainty avoidance ranking indicates the country has less concern about ambiguity and uncertainty and has more tolerance for a variety of opinions.3 (Hypothesis 4 will utilize this dimension)

Ø Long-term versus short-term orientation: values associated with long-term orientation are thrift and perseverance; values associated with short-term orientation are respect for tradition, fulfilling social obligations, and protecting one’s ‘face’. The values of this dimension are found in the teachings of Confucius, a high civil servant in China around the time of 500 B.C. Confucius’ teachings are lessons in practical ethics without any religious content.

Those five cultural dimensions identify core values that attempt to explain the general similarities and differences in cultures around the world.

The main characteristics of the first four dimensions are summarized in table 1 on the next page, which also helped to formulate the hypotheses (see chapter 4).

3

(11)

Table 1. Summary of the main characteristics of each cultural dimension

Score

Dimension Low High

Power distance index Minimum inequality Decentralization

Interdependence between powerful and less powerful people

Inequality accepted Centralization

Less powerful people have to deal with counterdependence

Individualism More collectivist nature with close ties between individuals Protect interest of group

Individuality and individual rights very important Less influenced by group norms

Masculinity Low level of differentiation

and discrimination between genders

More overlapping social roles for both men and women Neither gender needs to be overly ambitious or

competitive

Societies value qualities such as interpersonal relationships and concern for the weak

High degree of gender differentiation

Men are assertive, ambitious, competitive and strive for material success

Men respect whatever is big, strong and fast

Women care for nonmaterial quality of life, for children and the weak

Uncertainty avoidance index

Less concern about ambiguity and uncertainty

More tolerance for a variety of opinions

Little stress and fear Few rules preferred

Low tolerance for uncertainty and ambiguity

Much stress and fear Many rules preferred

Source: Hofstede (2005)

2.2.1 World Economic Forum background

(12)

Government.4 The annual meeting of top business leaders, national political leaders (presidents, prime ministers and others), and selected intellectuals and journalists is usually held in Davos, Switzerland, but there are also regional meetings throughout the year.

2.2.2 The status of women: current reality

The past three decades have observed a steadily increasing awareness of the need to empower women through measures to augment social, economic and political equity, to broaden access to fundamental human rights, and to improve nutrition, basic health and education. Gender refers to both women and men, and to their status, relative to each other. Gender equality refers to that stage of human social development when both men and women realize their full potential and opportunities, responsibilities and rights of individuals will not be determined by the fact of being born male or female. 5

2.2.3 Women and economy

The gender pay gap exists universally but its size might vary from one country to another. A variety of factors cause the gender pay gap but two important ones are occupational segregation and gender discrimination in labour markets. Gender discrimination occurs when people provide labour market services and are equally productive are unequally treated because of gender. Inequality means that these people receive different wages for the same work or face different demands for their labour services at a given wage.6 Worldwide, outside the agricultural sector, in both developed and developing countries, women are still averaging slightly less than 78% of the wages given to men for the same work, a gap which will not close in even the most developed countries.7 This means that for every US$1 earned by men, in both the industrial and services and manufacturing sectors, women’s earnings are in general only 78 cents.

While globalization has generated opportunities to reach international markets, it has in times intensified existing inequalities and insecurities for many poor women. They are usually the least able to seize the longer-term opportunities offered, since the gains of globalization are often

4_{Source: webpage: www.weforum.org}

5_{United Nations Office of the Special Advisor on Gender Issues} 6_{International Poverty Centre, UNDP June, 2006, Number 20} 7

(13)

concentrated in the hands of those with higher education.8 As shown in East Asia in the 1990s, it is all too often the case that women are only able to secure employment during rapid expansions.

The quality of women’s economic involvement is a particularly serious problem in developed countries, where women may gain employment with relative ease, but where their employment is concentrated in poorly paid or unskilled jobs. Upward mobility and opportunity is scarce. This is often the result of negative attitudes, and of legal and social systems that use maternity laws and benefits to penalize women economically for childbirth and child care responsibilities.

Internationally, women are most often concentrated in “feminized” professions, such as nursing, teaching, office work, and care. These functions are low paid, in addition to offering limited or no opportunity for advancement.

A vast majority of the world’s countries offer paid maternity leave, often with a guaranteed wage of 50-100% of salary. An interesting exception to this is the United States, who offers 12 weeks without any pay. A study has found 49% of high-achieving women to be childless, comparing to 19% of their male colleagues. Women who are in managerial positions often need to make a painful choice between a successful career and family.

2.2.4 Women and politics

Women´s low rate of participation at the highest levels of politics is an enduring problem in gender stratification. Despite advances in women´s levels of education and participation in the paid economy over the last 20 years, women have made little significant progress with respect to their representation in national politics. For example, in the United States, women compose 46% of the paid labor force and 55% of tertiary students9. However, their representation in the U.S. House of Representatives and the Senate remains 13% and 14% respectively. The situation is similar in other nations (Paxton, Kunovich, 2003). The Inter-Parliamentary Union (IPU) reports a world average of only 15,6% in combined houses of parliament. Other statistics by region range from 6,8% in the Arab States to 18,6% in the Americas, and 39,7% in the Nordic states.

8_{assuming the fact that higher education results in higher income.} 9

(14)

The absence of women from structures of governance inevitably means that national, regional, and local priorities are typically defined without meaningful input from women. Women have different life experiences and different awareness of the community’s needs, concerns and interests however. A study shows that when women should have more say in spending priorities they would be far more likely to spend on improving health, education, community infrastructure, and the eradication of poverty, as opposed to military, gambling and alcohol. In order to change those spending priorities, there must be a critical amount of women represented in decision-making positions.

Understanding women’s participation in politics begins by acknowledging that in no country do women compose 50% of the national legislature. There is a great variety across countries, from no women in countries such as the United Arab Emirates to 43% in Sweden. Previous cross-national studies on women in cross-national legislature have stressed three explanations for differences in women’s political representation: social-structural, political, and ideological (Kenworthy & Malami 1999; Paxton 1997; Reynolds 1999). Social-structural explanations focus on the pool of available women.

Availability supposed to cohere with educational and professional opportunities. Several researchers have tried to find empirical evidence for the correlation between women’s educational attainment and women’s representation. However, no cross-national study has found a statistically significant effect (Paxton, 2003).

Political explanations focus on the openness of the political system to women. It is generally accepted that the presence of a proportional-representation system, rather than a simple plurality system helps women in gaining access to the political system (Lovenduski & Hills 1981; Norris 1985; Rule 1987).10 An IPU survey proved, by 64% of the respondents, that proportional-representation systems are most conductive systems to the election of women. Other factors beside this proportional-representation system can play an important role, such as the presence of left-oriented political parties with an increase in women’s representation, national or party-level quotas, and the degree to which a country is democratic.

10

(15)

Ideological explanations focus on general impressions of women in politics and how viable women are as candidates and leaders. Despite the presence of favorable political systems or an adequate supply of qualified female candidates, cultural norms can play a pivotal role and can limit women’s opportunities to participate in politics. Public opinion about a women’s role and position in society can enhance or constrain women’s effort to seek political power. For centuries, political theorists such as Aristotle, Jean Jacque Roussea and others, justified the exclusion of women from politics because of their assumed nonrational nature. Despite women’s gaining of suffrage over the years proving the contradictory, this image is not totally disappeared so far.

2.2.5 Women and education

Educational attainment is the most fundamental prerequisite for empowering women in all fields of society. Women are not able to access well-paid, formal sector jobs, or participate in and be represented in government and other political functions without education of comparable level given to boys and men. Moreover, the next generation will risk low levels of education as well. It is very important to eradicate illiteracy since empirical work has shown that it reduces mortality rates of children and fertility rates.

The genderstats database of the World Bank gives some numbers on education and literacy.11 In the first place, public expenditure on education was 4% of GNI in 2004, an average number for all world countries. Female participation in education for primary education is 47% and for secondary education it is similar. However, when you consider countries separately, there are great differences. For example, in Egypt net enrolment rate (as percentage of age group) for secondary education is 81 % male and 77 % female for the year 2004. Statistics of 1990 show youth literacy rates for this country (as percentage of people aged 15-24); 70,9% male versus 51,0 % female. Data on Sub-Saharan Africa show a difference between male and female literacy for the year 1990 of 16 percent; 76,2% male versus 60,9% female.

Another important dimension, on which a knowledge gap between men and women has emerged, is information and communication technology. This has become a driving force of the development process. In countries where women have little access to information and

11

(16)

communication technologies because of lower education, countries are less able to use the benefits of the development progress for social and economic gain.

2.2.6 Women and health

According to a report written by the World Health Organization in 2005, from the 136 million pregnant women per year all over the world, 529.000 women die from causes related to pregnancy and childbirth (which is a percentage of 0.39). From the 46 million abortions worldwide, some 18 million are performed unsafely, which is more than a third. The result is that about 68.000 women die from complications, which is again a huge number.12

Another worrying fact about women’s health is Female Genital Mutilation (FGM). FGM is the partial or total removal of the female external genetalia.13 FGM is also called female circumcision. This wording sounds similar to male circumcision, however the degree of cutting is much more extensive, often impairing women’s sexual and reproductive functions (Toubia, 1993). Each year an estimated two million girls, usually aged 4 to 8, are subjected to FGM14. It is not uncommon that this female genital mutilation leads to death, chronic infection and bleeding, nerve tumours, obstructed childbirth and painful scarring. FGM is practiced globally. It is practiced in at least 26 of 43 African countries and in parts of India, Indonesia and Malysia as well (Toubia, 1993). In more developed countries FGM is also an important issue due to the continuation practice by immigrants from countries where FGM is common.

A third issue related to women’s health concerns violence against women. In a statement to the World Conference on Women in Beijing in September 1995, the UN Secretary-General Boutros Boutros-Ghali, said that “violence against women is a universal problem that must be universally condemned”. Women are particular vulnerable in three areas of concern: in the family (including domestic violence, traditional practices (FGM (explained above), son preference, and early marriage), in the community (including rape, sexual harassment, and prostitution) and by the State (including violence against women in detention as well as violence against in situations of armed conflict and against refugee women)15.

12_{World Health Organization (2005) “The World Health report. Make every mother and child count”}

13_{World Health Organization (1995). “Female Genital Mutilation: Report of a WHO technical working group”.} 14_{Amnesty International (2004)}

15

(17)

3. Literature review

3.1 Critical review on Hofstede

Geert Hofstede is the most-cited Dutch author and the ninth-most cited European in the Social Science Citation Index. Opions about Hofstede’s research are widely divergent. They vary from excellent and groundbreaking to very restrictive and confined. This section gives an overview of both kinds of comments. According to Bing (2004), Hofstede was someone who had analyzed the world of cultures through large-scale and quantitative research. Because of the scope of Hofstede’s research – over 115.000 questionnaires in over 50 countries – the findings had a very broad foundation. He calls Hofstede’s study groundbreaking for other reasons as well. Survey research had not been significantly used before in cross-cultural comparisons, certainly not across a large number of countries. Moreover, he claims that it is no exaggeration that Hofstede helped to create the field of comparative intercultural research.

Hofstede tries to defend himself to the abrasive criticisms of McSweeney (2002). There are five main criticisms that will be explained in this section and are summarized in table 2. below. McSweeney makes four crucial assumptions about Hofstede´s work and results. He argues that these results are all flawed and that therefore Hofstede’s national cultural descriptions are invalid and misleading. The first assumption is that every micro-location is typical of the national. Hofstede generalizes about the entire population in each country solely on the basis of analysis of a few questionnaire responses. A survey should not be the only way is Hofstede’s defense. However, Hofstede cannot understand how McSweeney could have read the book Culture’s consequences, without noticing the exploration of other cross-national differences. Exploration of other studies related to the IBM scores consist of cross-national survey and test data from other studies, including a number of representative samples of entire national populations, and of indicators measured at the country level such as GNP per capita, income inequality and percentage of the national budget of wealthy countries spent on development assistance to poorer countries.

(18)

questionnaire answers were significant; organizational, occupational and national cultures. Each respondent was conceived of as exclusively carrying or “permanently programmed” by these three non-interacting cultures. This reductive conception of IBM employees allowed Hofstede to argue that as there was only one IBM culture, the questionnaire response differences showed “national culture”. Hofstede specifically claims that what was measured were differences between national cultures.

Assumption four, that ‘identified’ in the workplace is unaffected by location, means that what is identified within a workplace is situationally non-specific. The IBM data were effectively restricted to the workplace. According to McSweeney, Hofstede ignores other sections of national populations, such as the unemployed, full-time students, retired, homeworkers etc. The questions were almost exclusively about workplace issues. The counterargument Hofstede gives falls again in the category of measuring differences between national cultures.

Table 2. Critique McSweeney, Answer Hofstede

Critique McSweeney Answer Hofstede

1 Surveys are not a suitable way of measuring cultural differences

They should not be the only way

2 Nations are not the best units for a studying cultures

They are usually the only kind of units available for comparison and better than nothing

3 A study of the subsidiaries of one company cannot provide information about entire national cultures

What was measured were differences between national cultures

4 The IBM data were old and therefore obsolete

The dimensions found are assumed to have centuries old-roots; only data which stayed stable across two subsequent surveys were maintained, and they have since been validated against all kinds of external measurements; recent replications show no loss of validity

5 Four or five dimensions are not enough Additional dimensions should be both conceptually and statistically independent from the five dimensions already defined and they should be validated by significant

correlations with conceptually related external measures.

Source: Author

(19)

McSweeney” gives answers on points not even treated by McSweeney. For example, ‘IBM data are old and therefore obsolete’, and ‘four or five dimensions are not enough’ are not given a chance in the article ‘a triumph of faith’.

However, several studies that used the Hofstede model, aimed at giving possible explanations for understanding. Harsh conclusions cannot often be drawn, but it can give at least directions for further understanding and further investigation. For example, a study done by Hofstede and Bond (1988) investigates the relationship between the fifth cultural dimension long-term orientation or Confucian Dynamism and Asian economic growth. The average annual sustained–growth percentage are high for the East Asian countries Singapore, Taiwan, South Korea, Hong Kong, and Japan. The countries of East Asia are said to have common cultural roots going far back into history. A test shows that the country scores on Confucian Dynamism derived from this exercise are strongly associated with those countries’ economic growth. Thus they have found a cultural link to an economic phenomenon. The values that compose the dimension of Confucian Dynamism do not seem to be recent developments caused by the fast economic development of certain countries. Therefore, the authors assume the values to be at least part of the cause, and economic growth to be the effect.

(20)

3.2 Gender inequality

There is a difference between the status of women in the developing and the developed world. In the poorest quartile of countries in 1990, only 5% of adult women had any secondary education, one half of the level for men. In the richest quartile, on the other hand, 51% of adult women had at least some secondary education, 88% of the level for men. Dollar and Gatti (1999) show that increases in per capita income lead to improvements in several measures of gender equality. They particularly address the fact that gender inequality in education is bad for economic growth. Thus societies that have a preference for not investing in girls, pay a price for it in terms of slower growth and reduced income.

The authors also mention an interesting point for my study. They are interested in what macroeconomic data reveal about three specific questions. If lower investment in girls’ education is an efficient economic choice?; does gender inequality reflect different social or cultural preferences?; and if there is evidence of market failures that may lead to under-investment in girls, failures that may decline as countries develop? If under-investment is efficient, then promoting girl’s education is a beneficial action. However, if it is not efficient but reflects cultural preferences, then promoting girls’ education may increase income, but will not increase a country’s welfare. When it is the result of market failure, then trying to increase girls’ education could be good anyway.

The paper gives evidence that girls’ education is not simply an efficient economic choice. Gender inequality in education can be explained by religious preference and other underlying characteristics of societies (think of culture). It also shows that gender inequality in secondary education is bad for growth and that societies have to pay a price for gender inequality in terms of slower growth. Moreover, the authors find evidence that increases in per capita income lead to reductions in gender inequality.

(21)

(22)

4. Hypotheses

The magnitude of the gender gap that exists in a country can have different causes. In my research, the gender gap index is composed of 5 dimensions and every dimension has different causes of the gap. For example, a low score on a country’s economic participation can be caused by high professional inequality and less chances in promotion. The score on the gender gap index can be interpreted the following way. A high score on the gender gap index indicates that a country has low gender inequalities and a low score on the gender gap index refers to strong gender inequalities.

However, for what reason it might be, in one culture certain circumstances are better accepted than in other cultures. This is why I would like to investigate all four cultural dimensions of Hofstede’s theory related to the overall gender gap of a country.

In total, six hypotheses will be tested. Four hypotheses are related to part one of my study ‘Culture and the gender gap’ and two hypotheses are related to part two of my study ‘The gender gap and economic development’. So firstly, I set up the following four hypotheses related to culture. The sign of the relationship is dependent on Hofstede’s theory and research.

Power distance, focuses on the degree of equality (or inequality), between people in a society. A high power distance ranking indicates that inequalities of power and wealth have been allowed to grow within a society, so I assume the differences between women and men to be better accepted. However, there can also exist a situation in which high power distance, characterized by a highly centralized legal system within a country and within businesses, will result in strict rules and well-organized systems, where the difference is not remarkable among genders. For example, unequal wage distribution (a characteristic of high power distance) does not need to be specifically present. Despite this possible situation, the first hypothesis will be:

Hypothesis 1: A country that has a high power distance score, will tend to have strong gender inequalities (so will score low on the gender gap index).

(23)

interdependence a society maintains among individuals. A low score on the individualism index corresponds to a collectivist society. I expect a collectivist society to care more about the overall well-being of the population of a country, so not only care about men, but as well as about the women in a society. However, caring does not mean per definition equal. For example, in Islamite countries where close ties between individuals are common, men and women are not often equally treated. Despite this fact, I assume that a collectivist country care more about women in comparison to individualistic countries, so I assume the following relationship. Hypothesis 2: A country that has a high individualism score, will tend to have high gender inequalities (so will score low on the gender gap index).

The following dimension, which is masculinity versus femininity, indicates the degree of gender differentiation. A high masculinity ranking indicates the country experiences a high degree of gender differentiation. A low masculinity ranking indicates the country has a low level of differentiation and discrimination between genders. In these cultures, females are treated equally to males in all aspects of the society. Since this dimension actually test the input for my research, the level of gender equality, I do not expect any contradictions among the sample, so the next hypothesis will be:

Hypothesis 3: A country that has a high masculinity score, will tend to have high gender inequalities (so will score low on the gender gap index).

(24)

of countries such as the United States, United Kingdom and Ireland. In those economies, the state plays an arm’s length role and labor has been progressively excluded. I expect the CME to be characterized by risk averse, or uncertainty avoidance (having more codes and rules), and women to be more involved in production. Therefore, the fourth hypothesis will be the following.

Hypotheses 4: A country that has a high uncertainty avoidance score, will tend to have low gender inequalities (so will score high on the gender gap index).

The following two hypotheses relate to the second part of my study ‘The gender gap and economic development’. In this second part of my study, I address issues of causality in two opposite directions. The causes and effects between gender inequality and economic development and in what direction. First, I will hypothesize the effect of economic development on strong or low gender inequalities. Second, I will hypothesize the effect of strong or low gender inequalities on economic development. The degree of association between the two variables has to be sufficient, and no other reasonable causes for the effect might be present.

However, you can also consider the causality the other way around. A country that has not the measures and input, which shows off in a country’s economic growth and development, is not able to close the gender gap sufficiently.

Hypothesis 5: A country with high economic development, will tend to have low gender inequalities (so will score high on the gender gap index).

A country that scores low on the gender gap index, that do not fully take advantage of one half of the talent in their population are misallocating their human resources and thus undermining their competitive potential. This will show off in a country’s economic development.

(25)

5. Methodology

5.1. Economic model and variables

What I have done for this research is combining two existing economic models. Moreover, I added an extra part to my research, to study growth and economic development. In this last part of my research I paid especially attention to the causality. The two existing models are Geert Hofstede’s cultural dimensions and the gender gap index of the World economic forum. I will explain the composition and construction of each of the two models below. In addition, this whole research makes use of secondary data.

Geert Hofstede’s cultural dimensions.

Hofstede´s primary data were extracted from a pre-existing bank of employee attitude surveys undertaken around 1967 and 1973 within IBM subsidiaries in 66 countries. He statistically analyzed the answers to these survey questions. That analysis revealed that there are four central and ‘largely dependent’ bi-polar dimensions of a national culture and that 40 out of the 66 countries in which the IBM subsidiaries were located could be given a comparative score on each of these four dimensions.

Ø Power distance index (PDI)

Ø Individualism versus Collectivism (IDV) Ø Masculinity versus Femininity (MAS) Ø Uncertainty avoidance index (UAI)

(For a broad description of each dimension, see section 2.1)

World Economic Forum report: “Women’s empowerment: Measuring the Global Gender Gap.” This study assesses the current size of the gender gap by measuring the extent to which women in 58 countries have achieved equality with men. Five important dimensions of female empowerment and opportunity have been chosen for examination, concerning global patterns of inequality between men and women:

(26)

The gender gap in each dimension is then quantified using two types of recent available data: A) published national statistics and data from international organizations, and B) survey data of a qualitative nature from the annual Executive Opinion Survey of the WEF16. Below follows a brief description of each of the five categories and the rationale behind them. In section 5.5, I will run a factor analysis to examine the underlying patterns or interrelationships for those variables and to determine whether the information can be condensed or summarized in a smaller set of factors or components.

5.1.1. Description of each of the five categories and the

rationale behind them.

Economic participation

The presence in the workforce in quantitative terms is an important step toward raising household income and encouraging economic development in countries as a whole.17 This view is strongly supported by a body of evidence suggesting that the education, employment and ownership rights of women have a powerful influence on their ability to control their environment.

The variable economic participation has been composed of data on unemployment levels, the levels of economic activity and remuneration for equal work.

Economic opportunity

Economic opportunity concerns the quality of women’s economic involvement. This is especially a serious problem in Third World countries where poorly paid jobs are characterized by the absence of upward mobility and opportunity. Legal and social systems which use maternity laws and benefits to penalize women economically for childbirth and child care responsibilities, and discourage men from sharing family responsibilities are in most cases the result of this.

For this variable, data is used on the duration of maternity leave, the percentage of wages paid during the covered period and the number of women in managerial positions. In addition, a

16_{WEF: World Economic Forum} 17

(27)

unique dataset on qualitative elements such as the availability of government-provided childcare, the impact of maternity laws on the hiring of women, and wage inequalities between men and women for private sector employment has included.

Political empowerment

This variable refers to the equitable representation of women in decision-making structures, both formal and informal, and their voice in the formulation of policies affecting their societies. National, regional and local priorities are typically defined without meaningful input from women as the absence of women from structures of governance shows.

Political empowerment has been measured by using data on the number of female ministers, seats in parliament held by women, women holding senior, legislative and managerial positions and the number of years a female has been head of state (president or prime minister).

Educational attainment

This is the most fundamental prerequisite for empowering women in all spheres of society. Without education of comparable quality given to boys and men, women are unable to access well-paid, formal sector jobs, participate in, and be represented in government and gain political influence. Moreover, society as a whole bears a higher risk that the next generation of children will be similarly low educated.

The variable educational attainment has been composed of data on literacy rates, enrolment rates for primary, secondary and tertiary education and average years of schooling across the population.

Health and well-being

Health and well-being refers to the substantial differences between women and men in their access to sufficient nutrition, healthcare, and reproductive facilities, and to issues of fundamental safety and integrity of person.

(28)

indicator of the lack of other choices available to young women. Finally, data on the percentage of births attended by skilled health staff, and maternal and infant mortality ratios are included.

The data used in this study come from publicly available sources, including the World Development Indicators of the World Bank, and the Human Development Report of UNDP, as well as the annual Executive Opinion Survey of the WEF18 (see for detailed operalization of data sources appendix 6). To make all data comparable, hard data was normalized to 1 to 7 scale, with the best value in each category being allotted a 7, and the worst value a 1. Once both survey and hard data are on the same scale, the unweighted mean of all variables within a particular category has been taken to calculate the scores for each country. Then, the overall scores for each country are calculated as an unweighted average of the scores obtained in each of the five categories.

The aim of this methodology is to provide cross-country comparisons. It shows easily the extent to which countries are making benefit of their full potential of their societies. This is why this study has attempted to consolidate several dimensions of gender equality in one index.

5.2 Research model

The research model I am going to use is a two-step research model. Firstly, I investigate the relationship between culture and gender, see figure 1 below. I expect that a country’s culture have an influence on the overall gender gap, and on each dimension of the gender gap separately. I will explain the methodology in further detail in the next section.

Figure 1. Research method – step one

Independent variables: Dependent variables:

18

In 2004, the Executive Opion Survey recorded the opions of nearly 9000 business leaders in 104 countries.

Hofstede’s cultural dimensions: Ø Power distance index (PDI) Ø Individualism versus

Collectivism (IDV) Ø Masculinity versus

Femininity (MAS)

Ø Uncertainty avoidance index (UAI)

§ Gender gap index (GGI): § Economic participation (EPA) § Economic opportunity (EOP) § Political empowerment (PEW) § Educational attainment (EDA)

(29)

Secondly, I would like to pay attention to the gender gap related to economic development (see figure 2 below). As the existence of the gender gap is already proven by the WEF rapport and has explained by country’s cultural differences, I think that it will be interesting to study whether there will be a relationship between this gender gap and economic development and in what direction it is influenced. In other words, the second research question is:

To what extent does economic growth foster gender equality, and to what extent does gender equality foster economic growth? Data on GNI and GDP growth rate are obtained from the following website www.unicef.org/infobycountry.

Figure 2. research method – step two

5.3 Sample selection

The analysis is conducted on a country level. The target population is all men and women living within a particular country. Purposive judgment sampling is applied because the Hofstede countries need to overlap with the countries used for the gender gap study. From the 80 countries in the Hofstede study, 43 countries were overlapping with the Gender gap index countries. Those 43 overlapping countries formed my sample.

I was only able to include countries that were used for those two studies. To give a judgment about selection bias, it is useful to consider their selection methods. Geert Hofstede selected his

§ Gender gap index (GGI): § Economic participation (EPA) § Economic opportunity (EOP) § Political empowerment (PEW) § Educational attainment (EDA)

§ Health and Well-being (HWB)

Economic development: Ø GNI

(30)

countries on the basis of the presence of IBM, from which he collected employee scores. The set of 58 countries covered in the current study includes all 30 OECD countries and 28 others from the “emerging market” world. Overall, the set of countries covers much of Europe and North America, in addition to providing relevant examples from Asia, Latin America, Africa, and the Middle East. I must admit that many countries from the developing world are omitting. This is due to lack of reliable data and this was the main consideration in the choice of countries.

The limitation of my study is that there are barely real poor and developing countries included.

The table in appendix 1 gives an overview of the country scores on Hofstede’s dimension for my study’s sample.

5.4 Regression analysis

I decided to use a multiple regression model, because I have more than one explanatory variable and because of the reasons mentioned above. This enables me to test a relationship between a dependent variable and more than one independent variable. I am going to use six regression equations in the following format:

Y = c + αA + βB + γC….+ ε

The independent variables will be six times the four cultural dimensions of Hofstede and the dependent variable will change. I first run the regression with the overall score of the gender gap (GGI) and then in turn use the components separately in the following order, economic participation (EPA), economic opportunity (EOP), political empowerment (PEW), educational attainment (EDA) and health and well-being (HWB). I add a constant to the equation, because this improves the accuracy in the regression mode. Also known as the intercept, it gives the value of the dependent variable when all of the independent variables have a value of zero.Table 6 on page 35 lists the signs I expect to come out of the equations.

(31)

One time with the gender gap index as dependent variable, and the GNI and GDP growth rate as independent variable.

After that, I do the tests the other way around, with the gender gap index as independent variable and the GNI and GDP as dependent variable.

5.5 Factor analysis

For my thesis, I make use of the gender gap index. This variable includes data on five other variables; Economic participation (EPA), Economic opportunity (EOP), Political empowerment (PEW), Educational Attainment (EDA), Health and Well-being (HWB). Factor analysis can be utilized to examine the underlying patterns or relationships for a large number of variables and to determine whether the information can be condensed or summarized in a smaller set of factors or components (Hair et al., 2006).

Many statistical methods are used to study the relationship between independent and dependent variables. Factor analysis is different in the way that it is used to study the patterns of relationship among many dependent variables. The goal is to discover something about the nature of the independent variables that affect them, even though those independent variables were not measured directly. Thus answers obtained by factor analysis are necessarily more hypothetical and tentative than is true when independent variables are observed directly. The inferred independent variables are called factors.

I think that a factor analysis is a useful way to discover patterns between the variables. For example, you can expect a country that ranks high on economic opportunity to rank high on economic participation as well. I run a factor analysis to identify underlying variables that explain the pattern of correlations within the five variables, which compose the overall gender gap index.

(32)

gender gap index. In addition, this one new component accounts for 47% of the variation in the original variables.

Table 4. Total variance explained

Initial eigenvalues

Factor Total % of variance Cumulative %

1 2.348 46.952 46.952

2 0.920 18.407 65.359

3 0.789 15.775 81.134

4 0.571 11.423 92.556

5 0.372 7.444 100.000

Table 5. Unrotated factor matrix

Measurement variables Component 1

EPA 0.738

EOP 0.550

PEW 0.789

EDA 0.727

HWB 0.591

A good solution should try to answer the question “How many components (factors) are needed to represent the variables?” My solution shows that only one component is being extracted. This means that the solution cannot be rotated. In my research I can considerably reduce the complexity of the data set by using only one component.

Considering the scree plot (see appendix 1), generally you want to extract the components on the steep slope. The components on the shallow slope contribute little to the solution.

The big drop occurs already between the first and second component, so using only the first component is an easy choice. I apply the latent root criterion of retaining factors with eigenvalues greater than 1.0, so one factor will be retained.

(33)

6. Regression analysis

The methodology part of this thesis specified six hypotheses that I test on a sample of 43 countries. My focus is on cross-country comparison. The way the hypotheses are set up suggests that correlation analysis could be used to discover the strength, the shape and the exact direction of the relationships. The objective of multiple regression is to predict the changes in the dependent variable in response to changes in the independent variables and that objective is most commonly achieved through the statistical tool of ordinary least squares (OLS). It estimates the coefficients α, β, γ… so that the sum of the squares (squared distance) between the actual and estimated value at any observation are minimal. The advantage of multiple regression is that all the coefficient estimates are simultaneously determined and the fact that R-squares give the overall model of fit.

Regression estimations support each coefficient with a significance level. Usually, three levels are considered to be acceptable: 99%, 95%, and 90% (probabilities of under 0.01, 0.05, and 0.1). Significance levels are very important regression indicators when testing hypotheses, because they tell us at what significance level the hypothesis is accepted or rejected.

Moreover, I run a factor analysis to discover underlying patterns or interrelationships between the five dimensions which compose the overall gender gap index. Despite facts that this factor analysis showed that these variables are pretty much related, it does not mean that running five separate regressions is useless. Running five separate regressions showed significant results. (see table. 9)

The level of prediction accuracy can be expressed by the coefficient of determination R².

(34)

point in time. Studies using time-series data usually have much higher R² values. Fortunately the success of a model cannot be completely judged on the magnitude of its R². Even if the numbers are low, the estimated parameters contain useful information.

6.1 Hypotheses testing and interpretation – part one

First, I will answer the hypotheses related to part one with a brief explanation. This will be followed up by an analysis of the obtained results. In section 6.3 and 6.4, I will do the same for part two of my study. In section 6.5, I give a summary of the findings and an overall interpretation of part one and two of this research. When possible, I explain how those two parts can be connected to each other.

Hypothesis 1: A country that has a high power distance score, will tend to have strong gender inequalities ( so will score low on the gender gap index).

Yes. It turns out that the higher power distance shows off in a country’s gender inequality. In countries with a high score on power distance, inequality is more accepted. Unfortunately, this fact is reflected in inequality among gender as well. The positive relation between higher power distance and high ranking on political empowerment can be explained as well. High power distance, more inequality accepted, and less ambition and therefore available political candidates among women, sounds as a logic explanation.

Hypothesis 2: A country that has a low individualism score, will tend to have low gender inequalities (so will score high on the gender gap index).

No. The relationship turned out to be the opposite. The more individualistic a country, the lower the gender inequality. A broad explanation for this is given in section 6.2.1., but it seems to be clear that individualism improves gender equality. Moreover, concerning the separate components, the results are contradictive as expected in table 6. This means that my outcomes are in total line with my finding for the overall gender gap index and hypothesis 2 will be rejected.

(35)

Yes. Actually this hypothesis tests the input for this research, the gender gap index. The dimension masculinity versus feminism indicates the level of gender differentiation within a country. My result is significant at a 99% level, so this hypotheses will be supported.

Hypotheses 4: A country that has a high uncertainty avoidance score, will tend to have low gender inequalities (so will score high on the gender gap index).

No. No association is being confirmed under this assumption. A country’s overall level of uncertainty avoidance has nothing to deal with gender inequalities. For example, Sweden has a low score on the uncertainty avoidance index (29) and ranks number one on the gender gap index. Another example, France, scores high on the uncertainty avoidance index (86) and high on the gender gap index as well (4,49). Those two countries show uncorrelated results.

6.2 Analysis of findings - part one

Table 6 below lists the signs I expect to come out of the equations. Table 6. Expected signs of regression coefficients – model 1

Dependent variables

GGI EPA EOP PEW EDA HWB

PDI - + + + + + IDV - + + + + + MAS - + + + + + Explanatory variables UAI + - - - - - Source: Author

Table 7 on the next page lists the estimated regression coefficients. Table 7. Estimated regression coefficients – model 1

Dependent Variables

GGI EPA EOP PEW EDA HWB

C 4,8256 15,8637 9,0068 6,9938 17,6307 12,2550 PDI -0,0103** -0,0337 0,0620 0,2139** 0,1465 0,1191 IDV 0,0114* -0,1004 -0,1381 -0,1617** -0,1665*** -0,1641*** MAS -0,0105* 0,2784* 0,1258 0,1219 0,1119 0,1205 UAI -0,0042 -0,0174 0,1543*** 0,0879 -0,0068 0,0858 R2 0,6077 0,2044 0,3201 0,5193 0,3002 0,3475 E xp la n a tory var iab les Adjusted R2 0,5665 0,1207 0,2485 0,4687 0,2266 0,2788 Source: Author

(36)

6.2.1 The overall gender gap explained by cultural

differences

The overall gender gap index (GGI) was regressed on the four cultural independent variables in the form of:

GGI = C + c(1) * PDI + c(2) * IDV + c(3) * MAS + c(4) * UAI

Table 7 shows the estimated regressions coefficients for this first model. Two of the four signs are as predicted. Power distance index is negative, which means that the higher the power distance within a country, the stronger the gender inequalities will be. In other words, the score on the gender gap will be low. The high power distance ranking indicates that inequalities of power and wealth have been allowed to grow within a society, so also between men and women. My sample supports this result, except for a few West-European countries. France has a quite high power distance score of 68 and a high gender gap index score of 4,49. Moreover, Spain, Belgium and Portugal show results that are not in line with my assumptions. France, as well as Portugal, score high on economic opportunity. Favorable legal and social systems, which includes maternity laws and wage equality, might be the reason for this.

Next, individualism is positive related to the gender gap index. My prediction was a negative relationship. This relationship has proved to be the opposite with statistical significance. According to the result, the more individualistic a country is, the higher the score on the gender gap index, or in other words, the lower the gender inequalities will be. For this positive relationship, I can think of the following explanations. Individual rights are very important and those refer to both men’s rights as well as women’s rights. For example, in the United States (score 91) a better quality of life and a high standard of living are aspired. This is considered a general belief among men as well as women. Considering the opposite, countries characterized by a collectivist culture such as Pakistan, Peru, Indonesia, Venezuela and Colombia, do not attain high gender equalities. Protection of interests of groups and caring more about the overall well-being of the population of a country do not symbolize a small gender gap. I think that in those more family-minded countries the gender roles are still more traditional.

(37)

ranking, will tend to have low gender inequalities, so will score high on the gender gap index. A low masculinity score indicates that a country has a low level of differentiation and discrimination between genders. Actually, this third Hofstede dimension test the degree of gender differentiation or gender inequality. I expected this relation to exist at a 99% significance level. Sweden clearly supports this result. It has a masculinity score 5 on this dimension, and ranks number one on the overall gender gap index, with an index number of 5,53. Other countries, such as Norway, Denmark and the Netherlands clearly support this relationship as well. There are no strong contradictions among the sample.

The last cultural dimension, uncertainty avoidance, does not show a significant relationship. There seems to be no association between risk adversity, or high uncertainty avoidance, and low gender inequality. Considering my assumption that coordinated market economies, including Sweden and Norway, prefer more codes and rules, so will be uncertainty avoidant, is not true. Sweden even has a lower number on the uncertainty avoidance index than the United States or the United Kingdom. However, Sweden ranks number one on the overall gender gap index, so taking this fact into account, no relationship between uncertainty avoidance and strong gender equality might be assumable.

(38)

6.2.2 Economic participation explained by cultural

differences

Economic participation (EPA) was regressed on the four cultural independent variables in the form of:

EPA= C + c(1) * PDI + c(2) * IDV + c(3) * MAS + c(4) * UAI

Economic participation refers to employment levels, the levels of economic activity and remuneration for equal work. The only variable, that shows to have a significant effect on a country’s economic participation, is masculinity. This result is in line with what was expected. The higher the degree of masculinity, the more a country appreciates men to be assertive, ambitious and competitive, as well as women to care for nonmaterial quality of life, for children and the weak. Women are supposed to participate less in the economy, which is indicated by a high ranking. However, there is one salient exception. China scores high on masculinity (66) and has a high economic participation. China is well-known for its highly forced driven economy, so this can be a possible reason.

6.2.3 Economic opportunity explained by cultural

differences

Economic opportunity (EOP) was regressed on the four cultural independent variables in the form of:

EOP= C + c(1) * PDI + c(2) * IDV + c(3) * MAS + c(4) * UAI

(39)

opportunity for women. In this regression, the relation between coordinated market economies, characterized by the fact that labor remains incorporated, can be reflected by higher economic opportunity also for women.

6.2.4 Political empowerment explained by cultural

differences

Political empowerment (PEW) was regressed on the four cultural independent variables in the form of:

PEW = C + c(1) * PDI + c(2) * IDV + c(3) * MAS + c(4) * UAI

Political empowerment explained by cultural differences shows again results as predicted. As seen in table 7, power distance and individualism have significant effects. To summarize, a higher a country’s power distance ranking, the higher a country’s ranking on political empowerment; and the higher a country’s level of individualism, the lower a country’s ranking on political empowerment. Political empowerment refers to the equitable representation of women in decision-making structures, and their voice in the formulation of policies affecting their societies. First of all, a high score on the power distance index means that inequality is accepted and centralization is common. It is not surprising that a country ranks high on political empowerment, which means a low representation of women in decision-making structures. This can be due to lack of available female candidates with political power. A study by Edmond Costantini (1990) documents the relatively low level of political ambition, defined in terms of a desire for political self-enhancement, among women. Second, individualism is characterized by importance of individual rights and less influenced by group norms. Equitable comparable representation among women in politics can be an effect from this in the sense that in individualistic countries honesty and public opinion are more appreciated, graduate degrees enhance one’s economic value or selfrespect, and work goes above personal relations (Hofstede, 2005). This holds for women as well, comparing to more collectivist countries.

Culture, Gender and Economic development

Msc. International Economics & Business

January, 2007