Rijksuniversiteit Groningen International Economics and Business

(1)

MASTER THESIS

CULTURE AND ECONOMIC DEVELOPMENT

Done by: Rouslan Gouriev (s1415042) Submitted to: Dr. Gabor Peli

Rijksuniversiteit Groningen International Economics and Business

12 September 2005

(2)

International Economics and Business

TABLE OF CONTENTS

ABSTRACT... 3

1. INTRODUCTION ... 3

2. LITERATURE REVIEW ... 5

2.1 Culture... 5

2.2 Culture, Economic Growth and Achievement Motivation ... 5

2.3 Change of Values... 6

2.4 Materialism and Postmaterialism... 8

2.5 Endogenous Growth Model ... 11

2.6 Relevance of the Study ... 13

3. THEORETICAL FRAMEWORK AND HYPOTHESES... 14

4. METHODOLOGY ... 17

4.1 Classification of the study... 17

4.2 Method of analysis... 18

4.3 Sample... 19

4.4 Data Sources ... 20

4.5 Variables ... 21

5. RESULTS ... 26

5.1 Descriptive Statistics... 26

5.2 Multicollinearity ... 28

5.3 Regression Analysis... 29

5.4 Regression Analysis for Developed and Less Developed Groups of Countries... 34

6. DISCUSSION... 37

7. CONCLUSION... 44

REFERENCES ... 45

APPENDIX A: List of Countries in 4-item Postmaterialism Index Sample ... 48

APPENDIX B: List of Countries in 12-item Postmaterialism Index Sample ... 49

APPENDIX C: List of Countries in Developed and Less Developed Countries Sample 50 APPENDIX D: Regression Models... 51

APPENDIX E: Calculation of 4-item Postmaterialism Index ... 55

APPENDIX F: Calculation of 12-item Postmaterialism Index ... 56

(3)

ABSTRACT

This study focuses on finding out how cultural dimensions (post-materialism and achievement motivation) and economic factors (income, investment, and primary and secondary education) affect economic growth across countries. The theoretical framework of this study is based on a study by Granato, Inglehart, Leblang (1996), who in their turn, use endogenous growth model developed by Levine and Renelt (1992). The main improvements of this study are the increased number of observations and use of 4 and 12 item indexes of post-materialism instead of only 4-item index used by Inglehart et al. (1996). The data for cultural dimensions is obtained from World Values Survey 1999- 2002. The main results of the regression analysis with 4-item postmaterialism index indicates that initial level of income, investment, and achievement motivation are significant variables when determining economic growth. However when 4-item index is replaced by 12-itme index initial level of income looses its significance and 12-item postmaterialism becomes significant and negatively correlated to economic growth.

1. INTRODUCTION

The question that will be posed in this paper is whether cultural factors influence economic growth across countries. If so, what is the best way to measure cultural dimensions, and what are their effects on economic growth? How strong is the influence of cultural factors compared to economic factors, and what is the nature of their relationship?

But, we must keep in mind that cultural factors alone do not influence economy of a nation, and there are variations in economic growth rates in each particular country throughout the time continuum. Also there are variations of economic growth rates across the countries. Granato, Inglehart, Leblang (1992) state that the rapid economical fluctuations that take a short time to change can be attributed only to economic factors, because cultural changes take longer time. But at the same time the power of culture to

(4)

change economic performance should not be underestimated. Well-known example of cultural influence on economic performance is a case of East Asia where Confucian- influenced economies managed to achieve a high level of economic development within a short period of time. However other nations, which shared similar endowments with countries of East Asia, and could take advantage of economic and political situation in that time did not manage to grow economically, therefore many scholars that study cultural effects on economies see it as a sign that cultural values indeed have an influence on economic development of the nations across the world.

There is a wide belief that a modern society enters the state of “post-materialism” values, where intellectuals, scientists, managers, dominate the political system, and profit making is not the highest priority within the nations (Inglehart, 1990). This theory explains the extend and nature of the influence that culture may bring to the society in terms of the economic development.

First section of this study elaborates on theoretical background of the chosen topic. This part emphasizes the importance of motivational factors exercised across the nations on the growth of economy of these nations. Post-materialistic values are also included in the analysis to further utilize the effect of different cultural attitudes on economic growth.

Second section introduces the data, which is based on World Value Surveys (1999-2002) (Inglehart et al. 2004). This survey allows me to construct indexes to measure the cultural dimensions – achievement motivation and post-materialism values. In addition I use the former 4-item index and improved 12-item index in order to accurately measure cultural values. To utilize the effects of economic factors on economic growth of the nations I use the baseline endogenous growth model presented by Levine and Renelt (1992). After necessary statistical analysis is conducted and robustness of economic-cultural model is estimated there is a section devoted to the discussion of the results and conclusions.

(5)

2. LITERATURE REVIEW

2.1 Culture

Culture is a very controversial topic in socio-economic studies. There is no single way to define culture. In general, it refers to human activity and different definitions of culture reflect different theories for understanding, or criteria for valuing, human activity.

Anthropologists use the term to refer to the universal human capacity to classify experiences, and to encode and communicate them symbolically.

Sir Edward B. Tylor wrote in 1871 that "culture or civilization, taken in its wide ethnographic sense, is that complex whole which includes knowledge, belief, art, morals, law, custom, and any other capabilities and habits acquired by man as a member of society" (www.wikipedia.com), while a 2002 document from the United Nations agency UNESCO states that culture is the "set of distinctive spiritual, material, intellectual and emotional features of society or a social group and that it encompasses, in addition to art and literature, lifestyles, ways of living together, value systems, traditions and beliefs".

Hofstede (1980) refers to culture as “the collective programming of the mind which distinguishes the members of one group from another and includes systems and values”.

Since values are determined earlier in life they tend to be programmed into individuals consistent with the cultural context and enduring all the time. Extensive physiological research shows the link between values and behaviors, and values are embedded in culture, Mueller and Thomas (2000) concluded that culture has an influence on economic performance of individuals and society as a whole.

2.2 Culture, Economic Growth and Achievement Motivation

The first approach that relates culture to economic development was proposed by Weber (1904) in his Protestant ethics thesis. His main argument is that rise of the Protestantism in Europe is a crucial event in the history of the economic development, because system of the protestant beliefs encourage norms, which are favorable to economic achievement.

(6)

This study was the first one to encourage the development of a whole stream of research that emphases achievement motivation with respect to economics. One of the most prominent scholars in this field is McClelland, his work in 1961 is built on the Weber’s thesis. However he specifically focuses on the need for achievement and defines it as a

“wanting to do well, with respect to standards of excellence”. He points out that different cultures look at the economic achievements differently; some emphasize it as a positive goal, others give it little emphasis. To support his claim he examines the values that were encouraged in children by their parents, school, and society as a whole. He measures emphasis in achievement through content analysis of the books, stories, and fairy tales used to educate children. He finds that some cultures emphasize the need for achievement more heavily and in these societies the economic growth is higher compared to the cultures where emphasis on achievement is not that strong.

At time McClelland findings were critisized based on the methodological grounds, mainly due to the fact that it was questioned whether he measured the values taught to children, or simply those that were written by writters. However during more recent study Granato et al. (1996), found the support for McClelland theory, by using the opinion survey in 25 countries around the world.

2.3 Change of Values

Throughout the time culture changes by implicit change of values that a given society exercises. Each culture represents people’s strategy for addaptation. In the long run this strategies generaly respond to economic, technological, and political changes. Cultures that fail to respond to the changes are unlikely to attain high economic development.

Though cultures change in response to changes in the socioeconomic, political and techological environment, they also shape the environment in return. Due to the changes in culture the industrial revolution in the West became possible and in its turn the industrial revolution brought a number of changes to the Western societies (Inglehart, 1990).

(7)

One of the theories that connects economic development and cultural change is the modernization theory. Its central claim is that economic development is linked with a coherent and to some extend predictable changes in culture and social and political life.

Evidence from around the world indicates that economic development tends to move societies in similar directions: industrialization leads to specialization, rising educational levels, rising income, change in gender rols, attitudes towards authority, sexual norms, declining fertility rates, broader political participation and less easily led publics.

However in contrast to socioeconomical changes, cultural changes are not similar across the countries (Inglehart and Backer, 2000).

Recently the research on socioeconomic developments gave rise to two schools of thought. One of them emphasizes the convergence of values as a result of modernization, that is economic and political changes will have significant impact on culture. This school predicts the substitution of traditional values with the modern ones. The second school emphasizes the persistence of the traditional values despite of the economic and political changes. They assume that values are relatively independent of the economic conditions (DiMaggio, 1996).

To test the validity of the two opposite theories Inglehart and Baker (2000) conducted a study of 65 societies. As a result they have found that economic development is associated with cultural changes and industrialization promotes a shift from traditional to more rational values, while the rise of postindustrial society brings a shift towards more trust, tolerance, well-being and postmaterialist values, hence Inglehart and Baker have found a partial support for the first school mentioned above. However they have also found that economic development tends to push societies in a common direction, but rather than converging, they seem to move on parallel trajectories shaped by their cultural heritage and path dependency plays significant role. They doubt that forces of modernization will produce one world culture.

On the other hand Wilensky (2003) when examining the structural and occupational changes on the societies criticizes the idea of materialization theory, concerning the

(8)

convergence of values and rise of postmaterialism values. He finds that even though structural changes with respect to organization at work and labour force trends are similar in general across the countries it is very difficult to identify a single model that can be prevalent in most of the countries. Especially when taking into account service sector which is dominant in developed countries but is too heterogeneous to draw any specific conclusions. With respect to the rise of post-materialism values Wilensky claims that these values can be applied only to a small proportion of population that consists of highly educated individuals and college students.

2.4 Materialism and Postmaterialism

Materialism has been defined as “the importance a consumer attaches to worldly possessions. At the highest level of materialism, such possessions assume a central place in a person’s life and are believed to provide the greatest sources of satisfaction and dissatisfaction” (Belk, 1985). There are three main views on the materialism. The first view was developed by Belk (1985) he viewed the materialism as collection of three personality traits: envy, non-generosity, and possessiveness. Envy is the “displeasure and ill will at the superiority of another person in happiness, success, reputation or the possession of anything desirable”. Possessiveness is “the inclination and tendency to retain control or ownership of one’s possessions”. Finally non-generosity is “an unwillingness to give possessions to or share possessions with others”. Latter Belk and Ger (1996) came up with a fourth trait – preservation, which is a tendency to make experiences tangible through souvenirs and photographs.

On the other hand Richins (1992) views materialism as a system of personal values. She defines materialism as a personal value stressing the importance of owning material possession. According to her classification materialism can be divided into three parts:

centrality, happiness, and success. Centrality is the general importance materialists attach to possession and the idea that possession plays a central role in their lives. Happiness is a believe that material possession will lead to well being and that possessing more or

(9)

better material things will improve one’s happiness. Finally according to Richins, materialists are people who believe that their success can be judged by the things they own.

The third view was developed by Inglehart (1990) based on Maslow’s hierarchy of needs, he also develops items to measure social values that reflect Maslow’s needs (1954). He sees materialism in a broader picture than simply behavior of the consumer. Inglehart defines materialist as somebody who sees the lower order needs for material comfort and physical safety, much more important than higher order needs such as: belonging, self actualization, self esteem, quality of life etc. Within this definition materialist would see strong economic growth, low crime rates and good institutional infrastructure to be most important social values.

According to Inglehart there is an opposite concept to materialism – postmaterialism. A postmaterialist is somebody who is more concerned with satisfying higher order needs, even at an expense of financial rewards. For postmaterialist the important social values would be: protecting freedom of speech, giving people more say in the community or work place and having society where ideas matter more then money.

There are a number of publications done by Inglehart on this topic. He started off with examining the change of values in postindustrial societies in his book “Silent Revolution:

Changing Values and Political Styles among Western Politics”. There he developed a theory of value change that is focused on specific changes that happened in the second half of the twentieth century such as: increasing education levels, decrease in the number of wars, technological renovations, economic changes and globalization. According to him these specific changes have led to the rise of the post materialist values in Western societies.

In his later publication “Culture Shift in Advanced Industrial Society” (1990) Ingelhart argues that there is a shift in political and social attitudes due to generational

(10)

replacement. In this publication he moves away from measuring postmaterialism on individual level to the society wide levels. This idea manifests itself in the next publication “Modernization and Post-modernization: Cultural, Economic, and Political Change in 43 Societies”. In this study he argues that postmaterialism becomes more common, which in turn brings a number of social changes, ranging from equal rights for women to establishing democratic institutions in the post-totalitarian societies.

The main argument that Ingelhart brings to public attention in his works is that modernization and industrialization has reached a point of diminishing marginal utility, making way for the shift from modern to post-modern society. In other words hierarchical institutions, bureaucracy, strict political, religious and cultural norms have outlived themselves. In addition the economic growth after WWII has stimulated the postmodern shift. Meaning that after the war was over, people could focus on satisfying higher order needs instead of focusing on daily survival. Increased financial security has led to the shift from materialistic to postmaterialistic values among age cohorts. Also Ingelhart made a point that post-materialists are not non-materialists or anti-materialists, but just individuals who more concerned with satisfying higher order needs then those of lower order.

De Graaf (1988) conducted a similar study in order to see the effects of the postmaterialism shift in the world. He has found that countries that rate high in postmaterialistic values have smaller emphasis on income attainment then materialistic countries, his findings are consistent with Ingelhart's description of postmaterialists as

“economic under achievers”. Also by conducting a research over a long period of time (1974-1985) de Graaf has found the support for the assumption that postmaterialistic value orientation within each particular culture is resistant to rapid change.

Besides doing the work on postmaterialism theory, Inglehart is also concerned with how culture influences economic growth and development. In 1996 Inglehart together with Granato and Leblang in their article “The Effect of Cultural Values on Economic

(11)

Development: Theory, Hypotheses, and Some Empirical Tests” examine how achievement motivation and postmaterialist values effect economic growth. Their first hypothesis state that achievement in a population is supposed to promote economic growth. By achievement in a population they mean a tendency of society to teach children such values as thrift and determination. The second hypothesis in that study assumed that prevalence of post-materialism values is supposed to have negative effect on economic growth. Granato et al. (1996) systematically measure cultural dimensions such as: achievement, motivation, and post-materialism on the sample of 25 countries that were part of “World Value Survey (1990-1993)”. Also Granato et al. (1996) used economic model of growth complimented with cultural values, so appropriate tests to examine cultural effects on growth are possible. They assess culture in the context of human and physical capital. In that study they ran several models that encompass both cultural and economic factors. In general the main implication is that growth rates are best understood as a consequence of both economic and cultural factors. Initially when they included only post-materialism and achievement, both of the variables turned out to be significant for economic growth, and had expected signs. However in the following model with both economic and cultural variables, only economic and achievement variables turned out to be significant. According to Granato et al (1996). This is probably due to the fact that countries with post-materialistic values are already fairly rich.

2.5 Endogenous Growth Model

Earlier in the history economic scholars developed two types of economic growth models, endogenous and exogenous growth models. The latter one was developed by school of neo-classic economists, among them Solow (1956) and Swan (1956). The main idea of this model is that growth is determined by exogenous variables, such as savings, population growth and technological progress, with a special attention given to the exogenous technological change. Later Romer (1989), Lucas (1989) and others decided to go beyond Solow’s model and attempted to develop models that explain technological progress itself. The result of their work was the development of an endogenous growth

(12)

model, one in which the long-run growth rate of output per worker is determined by variables within the model.

The choice of the economic variables in Granato, Inglehart and Leblang (1996) is based on the endogenous growth models that emphasize the role of the productivity of the population, initial income and savings. Also they work under the concept of “catching- up”, which means that the starting level of income is inversely related to the economic growth, in particular when countries are similar in technological parameters, then poor countries tend to growth faster, this concept was developed back in 1956 by Solow and Swan. In general endogenous growth models focus on a number of specific variables such as: initial level of wealth and investment in human capital, both of these variables are measured at the beginning of the period in for which the economic growth is measured. A number of studies among which Barro (1991) have found that these variables are significant when determining economic growth. Another set of variables used in endogenous growth models include: constant, physical capital investment and any variables that are of particular interest, such as culture in this case. The selection of economic variables are also underpinned by Levine and Renelt (1992) that finds that initial level of income, initial level of human capital investment and the period share of investment to GDP have significant correlation with the economic growth.

(13)

2.6 Relevance of the Study

This study is an extension of the study by Granato, Inglehart, Leblang (1996), where authors attempted to find how cultural and economic factors effect economic growth in a panel of 25 countries. I use the same framework, which is based on endogenous economic growth model used by Levine and Renelt (1992). The distinctive feature of my research is the fact that I extend the number of observations from 25 to 53 countries (Appendix A) and focus on a time period from 1980 to 2000 instead of 1960 to 1989 in Granato et al. (1996). Moreover in addition to four item index of post-materialism used by Granato et al. (1996) I use a twelve item index proposed later on by Inglehart in 1977.

By these modifications I plan to improve the fit of the regression, due to the increased variance of the data. My 53 countries sample includes both developed and developing nations around the world, whereas formal sample focused mostly on the developed countries. Furthermore a 12 item index may prove to be a better measure of postmaterialism.

Research Question: How do economic factors (income, investment, education) and culture factors (achievement motivation and post-materialism), influence economic growth across the countries?

Sub-Questions:

• How do economic factors (income, investment, education) influence the economic growth?

• How do cultural factors (achievement motivation and post-materialism) influence economic growth?

• Which of the two sets of variables have a stronger influence on economic growth?

(14)

3. THEORETICAL FRAMEWORK AND HYPOTHESES

This study is based on the framework of Granato, Inglehart, and Leblang (1996), which used endogenous growth model in order to examine the effects of economic and cultural variables on the economic growth in 25 nations for the period of 1960-1989. This model consists of two set of factors: economic and social or cultural factors. Whereas economic factors are mostly fixed, a set of social and cultural factors vary from research to research depending upon the interests of the researcher. The general form of endogenous growth model is as follows:

Yi = βIi,o + ΠXi + εi

Where Yi is output growth per capita for country i. Ii,o is a set of economic variables for country i measured in the beginning of the period in question. These variables include initial income per capita and investment in human capital (education), these variables are also measured in the beginning of a period in question. Xi is the set of other variables that includes constant, investment and other variables that researcher is interested in exploring. In this study these are postmaterialism and achievement.

I begin with endogenous growth model. According to this model the first concept to be used in this research is an initial level of income per capita. There are several studies like Barro (1991), Helliwell (1994), Levine and Renelt (1992) that have found a negative relationship between the initial level of income per capita and economic growth. The main argument behind this finding is the convergence in income levels across countries.

In particular if two countries are similar with respect to endowments the poorer country is expected to grow more during a certain period of time compared to the richer country.

Poor countries where ratio of capital to labor is low have high marginal products of capital and therefore tend to grow at higher rates. Moreover international capital mobility and technological innovations have positive effect on a rate of economic growth. Based on this evidence first hypothesis is as follows:

(15)

H1: Higher initial level of income per capita will lead to lower rate of economic growth.

Second concept is the investment in human capital, which is usually measured in education. For example Chevalier, et al. (2003) attempted to study effects of education on productivity of the individuals. In the study they managed to find a relationship between presence of higher education and productivity of individual, in comparison with individual who does not have higher education. It seems that people with education tend to have a better return in terms of their productivity. Another study done by Wolff (1999) examined the effects of education on economic growth within 24 OECD countries. His main finding was that there is a positive relationship between formal education and economic growth. Especially higher education is increasingly important, for economic growth. More specifically he found that not a number of school years, but a number of scientists and engineers is important. This evidence makes it possible for me to assume that education is indeed a significant variable in determining economic growth. Hence my next hypotheses:

H2: Higher percentage of individuals enrolled in primary education will lead to higher economic growth.

H3: Higher percentage of individuals enrolled in secondary education will lead to higher economic growth.

The fourth concept is the physical capital investment. Both Barro (1991) and Levine and Renelt (1992) found that increasing rate of physical capital accumulation increases a nation’s rate of economic growth. Economic growth, that is an increase in per capita income, can be achieved by bringing more resources into productive use or by deriving more output from a given amount of resources. The concept of the creation of physical and human capital is to produce a good without intending to directly consume it. It is intended to increase the level of productivity of a nation, allowing for an increase in the future production of a nation. It is believed that creation and accumulation of capital

(16)

depends on the ability of a nation to give up consumption in a current time, in order to enjoy the benefits of capital accumulation later. Hence the next hypothesis is:

H4: The higher levels of physical capital investment will lead to higher economic growth.

The fifth concept and the first “cultural” variable is achievement motivation. This concept was mentioned by Granato, Inglehart, and Leblang (1996), where they found a significant positive correlation between the emphasis on achievement in a particular country and its economic growth. Their findings were routed in Weber thesis that some societies pay more attention to the achievement then the others and bring up values such as thrift and determination in their children. Based on these works it is possible to assume that there is a positive relationship between achievement and economic growth:

H5: Countries with cultures oriented towards achievement will have higher economic growth.

The final concept in this research is postmaterialism. According to Granato, Inglehart, and Leblang (1996) postmaterialistic values are predicted to have negative effect on economic growth, since postmaterialist are mainly concerned with higher order needs such as self-actualization, self-esteem, and belonginess instead of lower order needs which are more associated with production of goods, hence economic development.

Moreover there is evidence that postmaterialist values emerge when society attains a rather high level of economic security and people can afford to spend their time on something else besides work. And as it was mentioned before developed nations might have a lower rate of economic growth compared to developing nations (given similar endowments), where people are more likely to have materialistic values due to economic instability.

H6: Countries with cultures oriented towards postmaterialism will have lower economic growth.

(17)

4. METHODOLOGY

4.1 Classification of the study

The proposed research is applied since it is problem solving and trying to provide an answer to the specific research question. Given research is the best classified to be of explanatory nature, as its main aim is to explain how economic and culture mentioned above influence the economic growth across the countries. After conducting the statistical analysis in form of regression, the research will be ended by the discussion and conclusion, where relationships between dependent and independent variables will be elaborated in details. Concerning the degree to which research question is crystallized my study is formal, as it has the hypothesis, involves precise procedures and uses specific data sources. Moreover this study can be used further in the research as a basis for the development of other testable hypothesis.

Next issue is the method of data collection, it falls under monitoring, as research does not use any questioners or interviews, but focuses entirely on the secondary data, which is available in the literature and on Internet. Since I use only secondary data I do not have any influence on the variables, so I can only report on what has happened. In this case the research can be classified as of ex post facto design. Concerning the purpose of the study, it can be classified as causal, because it tries to explain the relationship between variables. The time dimension of the study can be classified to be cross-sectional.

However the dependent variable is the average number of the GDP growth through out the years, all other variables represent some point in time. The topical scope of this study falls into dimensions of statistical study, because statistical techniques will be used to find out the relationship between the given variable. The statistical studies try to capture population characteristics making inferences form the sample characteristics. The hypotheses are tested empirically and generalizations about findings are presented based on representatives of the sample and validity of the design.

(18)

4.2 Method of analysis

For this research the method of analysis that I choose to use is regression. Regression analysis is used to test statistical models of the relationship between variables. Regression analysis implies that there is a dependent variable and one or more independent variable.

The variable being predicted or explained by the equation is called the dependent variable and the variable being used to predict or explain the dependent variable is called the independent variable. A multiple regression allows incorporating more than one independent variable into an equation. In this case the estimating equation may look like this: y=β0+β1x1+β2x2+β3x3. This type of analysis is useful in two ways. First, it almost inevitable offers a fuller explanation of the dependent variable, since few phenomena are products of a single cause. Secondly the effect of a particular independent variable is made more certain, for the possibility of distorting influences from the other independent variable is removed.

There are few assumptions that are needed for the regression analysis to be feasible:

• The assumption of normality: all variables have normal distributions

• The assumption of reliability: all variables are measured accurately, and there is no correlation between variables. Also values of y are independent of each other.

• The assumption of Homoscedasticity: that is the variables have constant variance.

The problem is to find values for β0 and β1,2,3.. such that the line described by the equation is the line of best fit. One method is to use the technique of ordinary least squares. This is an example of something called a maximum likelihood method. In the least squares method the best-fit line is defined as that line which minimizes the sum of the squared differences between actual and predicted values of y, i.e. it is a sum of squares method. There are several tests to determine the significance of the relationship between the variables. The F test is used to determine whether significant relationship exists between the dependent variable and the set of all independent variables, this is a

(19)

test for overall significance. A separate T test is conducted for each of the independent variable in the model, this is an individual significance test. (Anderson et al, 2002).

One of the main problems with a multiple regression analysis is the possibility of multicollinearity, meaning that one or more independent variables may be correlated with each other. When multicollinearity exists it is impossible to arrive at a unique solution for the least squares parameter estimates. The general difficulty is that parameter estimates become unreliable. Hence there is little confidence that a particular slope estimate accurately reflects the impact of X on Y in the population (Lewis-Beck, 1993).

4.3 Sample

My sample consists of 53 observations (Appendix A). Based on this sample I expect to draw the conclusions about the entire population. In order for sample to be representative it should have approximately the same distribution of characteristics as the whole population has. The best way to obtain the representative sample is by probability sampling, which is a technique that depends on the mathematical probability that a number of members carefully selected from a larger group will be more or less representative of that group. Unfortunately due to data and time limitations I am not able to use this technique and my sample is obtained by non-probability method. That is why this is a random sample, which I had a chance to come across. However it is still possible to use 53 country sample in this research. First of all the sample is obtained from reliable statistical sources. Secondly it has been used in previous studies, in addition my sample includes a higher number of observations (53 instead of 25). And finally the sample is varied enough, since it represents countries that vary in their economic development and cultural and institutional settings. Also I focus on 20 years from 1980 to 2000. My data set of 53 countries includes all 25 countries used in the previous research by Granato et al. (1996) (Appendix A), hence there are no missing countries and 28 new ones are added. In addition to the developments mentioned above there is one more set of countries which are used in this research when testing the influence of 12-itme

(20)

postmaterialism index on the rate of economic growth. This data set has 37 observations (Appendix B), 12-item index was not used before by Granato et al. (1996).

In addition to the to the regression analysis with full 53 countries data set, I perform an analysis within two groups of countries: developed and less developed. The reason for doing division is the fact that economic growth measured in percent is a relative concept and high economic growth in one country does not necessarily mean that this country is richer compared with countries with lower economic growth. "Less developed country"

classification employed by the United Nations includes all of Africa, all of Asia except Japan, the Transcaucasian and Central Asian republics of the NIS, all of Latin America and the Caribbean, and all of Oceania except Australia and New Zealand (U.S. Census Bureau, 1996). In addition to geographical division I divided countries by their GDP, countries with GDP per capita in 1980 above 10.000 USD are considered to be developed and countries with GDP per capita 1980 below this number are less developed. In total there are 24 developed countries and 29 less developed countries. (Appendix C).

4.4 Data Sources

Data on culture variables is obtained from the book written by Inglehart, Basanez, et al.

“Human Beliefs and Values” (2004). This book includes data on basic values and attitudes of the people in 82 countries around the world. Based on the 1999-2002 World Values Survey, it provided standardized cross-cultural measures of people’s values and goals, concerning politics, economics, culture, religion, and others aspects of society. The total number of observations present in this study comes up to 200,000. The data provided in this book is used to calculate indexes on achievement motivation and postmaterialism. I choose to use only 53 countries out of 82, because it was possible to calculate culture indexes and obtain full data on other variables only for 53 countries.

(21)

The data on economic variables such as: GDP growth per capita, GDP per capita, and primary and secondary school enrolment is obtained from World Bank's The World Development Indicators (WDI). Finally data on investment is obtained from Penn World Table Version 6.1, Center for International Comparisons at the University of Pennsylvania (CICUP).

4.5 Variables

Dependent variable: Mean rate of Per Capita Economic Growth (1980-2000)

This variable is calculated from annual percentage growth rate of GDP per capita, which is based on constant local currency. GDP per capita is a gross domestic product divided by midyear population. GDP at purchaser's prices is the sum of gross value added by all resident producers in an economy plus any product taxes and minus any subsidies not included in the value of the products. It is calculated without making deductions for depreciation of fabricated assets or for depletion and degradation of natural resources. To obtain the mean rate of per capita economic growth, the sum of annual percentage growth rates of GDP per capita is divided by the number of years observed.

Independent variable: Per Capita GDP

According to the endogenous economic growth model GDP per capita should be taken only for the base year (1980). The variable is obtained from World Bank's The World Development Indicators, and represents real per capita GDP with 1995 for a base year.

Independent variable: Primary and Secondary Education

This variable takes on the form of a gross enrollment ratio, and is taken only for the base year (1980). Gross enrollment ratio is the ratio of total enrollment, regardless of age, to the population of the age group that officially corresponds to the level of education shown. Primary education provides children with basic reading, writing, and mathematics

(22)

skills along with an elementary understanding of such subjects as history, geography, natural science, social science, art, and music.

Secondary education is a period of education, which follows directly after primary education, and which may be followed by tertiary or "post-secondary" education. The purpose of a secondary education can be to prepare for either higher education or vocational training. The exact boundary between primary and secondary education varies from country to country and even within them, but is generally around the seventh to the tenth year of education, with middle school covering any gaps. Secondary education occurs mainly during the teenage years.

Independent variable: Investment

This variable is measured as an average for the period1980-2000 of the ratio of the real domestic investment to real GDP which is measured from the base year of 1996.

Independent variable: Achievement Motivation

In order to calculate achievement motivation index World Values Survey (1999-2002) is used. This survey asks representative national samples the following question: “Here is a list of qualities which children can be encouraged to learn at home. Which of them if any do you consider to be especially important?”. According to Inglehart, Granato and Leblang (1996) the list of offered answers includes two items that can be considered as values emphasizing achievement, such as “thrift, saving money and things” and

“determination”. Two other items reflect emphasis on the conformity to the traditional social norms, such as “obedience” and “religious faith”. The index of achievement motivation is constructed for each country by summing-up the percentages emphasizing the first two goals minus the percentage emphasizing the latter two goals. This method of index construction allows identifying countries that place heavy emphasis on the achievement compared to countries that do not put such a heavy emphasis on achievement.

(23)

Independent variable: Postmaterialism

Originally the post-materialism index was developed by Inglehart back 1971, it was aimed to measure the extend to which a society adheres to postmaterialistic or materialistic attitudes. It was based on four items measuring an individual’s hierarchy of policy relevant issues. This index had been used since 1970 in the Eurobarometer-series and since early 80’s in the World Value Survey. The index was based on the following question:

“People sometimes talk about what the aims of this country should be for the next ten years. On this card are listed some of the goals which different people would give top priority. Would you please say which one of these you, yourself, consider the most important? And which would be the second most important?”

1^st Choice 2^nd Choice 1. Maintaining order in the nation 1 1

2. Giving people more say in 2 2 important government decisions

3. Fighting rising prices 3 3

4. Protecting freedom of speech 4 4

Respondents were asked to choose the most important one first and the second important one second. Those who’s first choice was item “1” and the second choice “3” or the first choice “3” and the second “1” were rated as materialists. Those whose first preference was item “2” and the second choice was “4” or the other way around, were rated as postmaterialists. Later Inglehart introduces the category “mixed type” which is seen as a person being able to swing to other side by choosing only one item from “1” and “3” and the other from “2” and “4” (Appendix D).

The Inglehart’s index had received a lot of criticism throughout the years, mainly because it was thought to reflect only political attitudes. For example Klein (1995) based on his study of Germany from 1973 to 1992 criticizes the one dimensional way of measuring the value change. He runs a multinomial-logit model according to age cohorts to examine

(24)

when individual changes his or her values. He concludes that there is only a relative change and there is no general linear strand towards post-materialism. He also states that materialistic orientations are due to generational settings. And more postmaterialist orientations take place after reaching physical security and it is due to the socialization process. Also Clarke et al. (1999) studies how economic influences affect the index. He argues that index only measures current political issue and does not account for long-term value change.

In response to critics Inglehart develops a 12 item index. Which was designed to permit a fuller exploration of Maslow’s need hierarchy. Moreover this index allows to capture a long term preoccupation of individuals instead of one’s response to immediate situation.

Six out of 12 items in the index are designed to utilize materialist value priorities:

fighting inflation, promoting economic growth, maintaining economic stability, maintaining order, combating increasing levels of crime, and maintain strong defense forces. Other six items were designed to utilize post-materialist priorities: more say in government, free speech, more say in the work place, beautified cities and countryside, humane society, and society where ideas count (Appendix E). In contrast to 4-item index the 12-item can help to reduce the errors and measurements and provide better fit for a regression. Consequently in my research I use both indexes.

(25)

So the final equation has the following form:

y=β0+β¹x1+β²x2+β³x3+β⁴x4+β⁵x5+β⁶x6+ε Where:

y Dependent variable: Mean Rate of Per Capita Economic Growth

β0 Constant

β¹,β²,β³,β⁴,β⁵,β⁶ Coefficients

X1 Independent variables: Per Capita GDP (1980)

X2 Primary Education

X3 Secondary Education

X4 Investment

X5 Achievement Motivation

X6 Post-materialism (4 and 12 item)

ε Error term

(26)

5. RESULTS

5.1 Descriptive Statistics

Data set used in this research consists of 53 observations. The dependent variable is growth rate of real per capita GDP from 1980 till 2000. The lowest growth rate is - 1.4690% observed in Venezuela and the highest value observed in China reaching 8.1636%. The mean value of this variable is 1.7981% with a standard deviation of 1.6360% (Table 1).

In addition to the dependent variable there are seven independent variables, which can be divided into two groups, economic and social. The first economic variable is the 1980 value of real per capita GDP with 1995 as a base year. For this variable there are 53 observations. The lowest value is 113.92 USD observed in Morocco, and the highest value of 39,841.75 USD can be found in Switzerland. The mean value of this variable is 9,682.19 USD with a standard deviation of 9,830.84 USD. The next independent variable is number of students enrolled in a primary school grade level relative to the total population of that age group in 1980. The lowest number of students in primary education is found in Pakistan (40.2133%) and highest 123.2378% found in Portugal.

The mean value for this variable is 97.5769% with standard deviation of 15.6063%. The third independent variable is secondary education, defined by a number of students enrolled in secondary school grade level relative to the total population of that age group in 1980. The minimum value observed for this variable is 3.3021% found in Tanzania and maximum value for this variable is 104.8116% observed in Denmark. For this variable one observation is missing and the mean value is 60.1164% with a standard deviation of 28.7886. The fourth independent variable used in this research is average ratio of real domestic investment to real GDP from 1980-2000. The minimum and maximum values observed for this variable are 2.8139 and 44.1376 found in Uganda and

(27)

Singapore respectively. There are all 53 observations present for this variable with a mean value of 19.2339 and standard deviation of 7.3671.

The first social variable and fifth independent variable in regression is four item achievement motivation indexes. With the lowest value of -1.23 observed in Egypt and highest value of 2.13 observed in Austria. The mean value for this variable is -0.05 with a standard deviation of 0.57. The next two social variables refer to postmaterialism and are used in two separate regressions. Both values represent the same concept, but are calculated differently. In the first case postmaterialism index is based on 4-item battery and the second case postmaterialism is measured by the 12-item battery. For the 4-item postmaterialism index the minimum and maximum values observed are 1.42 and 2.72 found in Pakistan and Belgium respectively and the mean value of 1.86 and standard deviation of 0.22927. For the 12-item postmaterialism index the minimum and maximum values observed are 1.10 and 2.80 found in Albania and Switzerland respectively and the mean value of 2.03 and standard deviation of 0.45, unfortunately 12-item postmaterialism is available only for 37 countries due to the lack of data.

Table 1. Descriptive Statistics

N Minimum Maximum Mean Std. Deviation

Grothpercap 53 -1.4690 8.1636 1.798084 1.6360358

GDP1980 53 113.92 39841.75 9682.1998 9830.83811

PrimEd 53 40.2133 123.2378 97.576947 15.6062636

SecEd 52 3.3021 104.8116 61.116475 28.7886187

Invest 53 2.8139 44.1376 19.233928 7.3670788

Achievment 53 -1.23 2.13 -.0526 .56523

Postmaterialism 53 1.42 2.72 1.8630 .22927

Postmaterialism12 37 1.10 2.80 2.0338 .45009

Valid N (listwise) 36

(28)

5.2 Multicollinearity

One of the main issues in the multiple regression analysis is a possibility of a correlation between independent variables. When multicollinearity exists the standard errors tend to be higher and this in turn results in smaller values for t ratios, so there is a possibility of the situation that F-test shows that the overall relationship is significant but none of the coefficients is significant. As a result the estimated regression equation provides poor insight into the relationship between given variables. There are several ways to identify multicollinearity; one of them is by inspection of a matrix of bivariate correlations. As a rule of thumb a correlation is considered to be significant if the correlation coefficient between a pair of independent variables exceeds 0.7. However this test has its drawbacks, more specifically a severe multicollinearity problem may not be reflected in these correlations: one independent variables may be approximately a linear combination of several other independent variables in a model yet that variable might not be highly correlated which each particular independent variable. Secondly due to variety of sample sizes it is sometimes rather difficult to define an appropriate cut-off value. A preferable test for multicollinearity is to regress each independent variable in the model on all other independent variables and look for R² values that are close to one.

Table 2. Correlations

Variables GDP1980 Investment Primary Education

Secondary Education

Achievemen t

Postmaterial ism

R²

GDP1980 1 0.769

Investment 0,641 1 0.764

Primary Education

0.162 0.435 1 0.350

Secondary Education

0.801 0.632 0.360 1 0.717

Achievement 0.615 0.569 0.275 0.623 1 0.461

Postmaterialism 0.583 0.288 0.274 0.543 0.320 1 0.437

(29)

In the given regression model there are several correlations that are worth mentioning.

First of all correlation coefficient shows that secondary education is correlated to GDP from 1980. The values of R² also shows several correlations, although they are not perfect R² has rather high values of more then 0.7. The high correlation can be observed in a case of GDP 1980, Investment and Secondary Education. Unfortunately even though there are tests to identify multicollinearity there are no reliable techniques to identify the degree to which multicollinearity affects the results. One way to deal with multicollinearity is to obtain more information on a variable, however there are no missing observations for these variables in my data set. Another way to deal with this problem is to use prior knowledge about the relationship between independent variables.

Finally it is possible to exclude variable that shows correlation from the model, but in this case it will undermine the theoretical framework behind the model, and since I base my selection of variables on prior research and theoretical models I cannot exclude any variables without altering the model. Considering these options I choose to leave out the correlated variables from the model, and keep the existing correlations in mind, while interpreting the results.

5.3 Regression Analysis

For the regression analysis I follow the framework used by Granato et al. (1996), they estimate several models that firstly include only economic variables, then only social variables and finally all of the variables grouped together in order to identify only significant ones.

(30)

Table 3. OLS Estimation of Economic Growth Models. Dependent Variable: Mean Rate of Per Capita Economic Growth (1980-2000)

Model Variable Model 1 Model 2 Model 3 Model 4 Model 5 Constant 0.858

(1.313) 6.014*

(1.808) 3.787

(2.129) 0.710

(0.678) 4.486*

(1.454) Per Capita GDP in 1980 -9.945E-05*

(0.000) -9.373E-05*

(0.000) -8.532E-05*

(0.000) -5.499E-05 (0.000) Primary education 1980 -0.018

(0.016)

-0.015 (0.016)

Secondary education 1980 0.017

(0.013) 0.013

(0.013) Investment 0.136*

(0.041) 0.106*

(0.042) 0.102*

(0.037) 0.096*

(0.038) Achievement Motivation 1.200*

(0.390) 0.988*

(0.481) 1.063*

(0.468) 1.877*

(0.585) Postmaterialism 4-item -2.229*

(0.961)

-1.326 (1.147)

Postmaterialism 12-item -1.832*

(0.733)

R² 0.229 0.187 0.318 0.269 0.496

R² adjusted 0.165 0.154 0.229 0.224 0.433

SEE 1.495 1.505 1.437 1.441 1.408

F 3.571 5.740 3.575 6.000 7.866

Cook Distance (max)

0.233 0.652 0.345 0.456 0.402 Notes: Standard errors are in parentheses

* T-test significant at 95% confidence level.

First I estimate the endogenous growth model that regresses the nation’s rate of per capita growth on its initial levels of per capita income, education, and the rate of the physical capital accumulation. The first test to conduct when making the regression is the F test, its purpose is to test whether a significant relationship exists between the dependent variable and the set of independent variables. This is the test for the overall significance, the F-test tests two hypotheses: one of them is that one or more of the parameters of the multiple regression model is equal to zero, the alternative hypothesis is that one or more parameters is not equal to zero. The first hypothesis is rejected if calculated F-value is higher then the critical value. I conduct this test at 95% confidence level meaning that I

(31)

can reject or accept given hypothesis with 95% confidence. With 95% confidence level and 4 and 48 degrees of freedom in numerator and denominator respectively, the critical value of F is approximately 3.10. In this regression the calculated F value is 3.571 (Table 3) which is more than the critical value. This result gives the possibility to conclude that overall relationship between dependent and set of the independent variables is significant.

The following value that is essential in the regression analysis is R² or multiple coefficient of determination. It can be interpreted as the proportion of the variability in the dependent variable that can be explain by estimated multiple regression equation. The rule of thumb is that the higher the value of R² the better dependent variable is explained by independent variables. In the given regression analysis the value of R² is 0.229 or in another words the estimated regression equation explains 22.9% of the variability in the independent variable. R² can also be adjusted for a number of independent variables, and this value gives better prediction because it is more reliable then R², usually adjusted R² is smaller in value then R². In the given regression R² adjusted is 0.165% (Table 3).

Another problem in multiple regressions is significant outliers, that is the values that are unusually large or small. Outliers may influence overall results of regression, their presence may be a signal that the regression model fails to capture important characteristics of the data. One way to identify outliers is to calculate residuals for each observation, however it is useful to have a summary index of the influence of each observation on the fitted regression line. To identify such influential observations, Cook’s Distance is used. As a rule of thumb values of Cook Distance for each particular observation that are higher then 1 are considered to be influential outliers. In the given regression analysis the maximum value is 0.233 (Table 3), which is significantly smaller then 1, meaning that there are no evidence of influential outliers.

The next step is to identify which independent variables are significant by means of t-test.

T-test is referred as a test for individual significance. It proves the validity of the Ho hypothesis, that each particular coefficient is equal to zero. If calculated value of t is

(32)

higher then the critical value, or lower in case of negative coefficients, then Ho is disapproved and alternative hypothesis holds, meaning that each particular coefficient is not equal to zero. Table 3 does not show the values of t but only indicates which variables are significant. Appendix C provides the exact values of t for each particular variable. At 95% confidence level and 48 degrees of freedom the critical value is 2.021, that is if t- value for each variable is higher then the critical value, or lower in case with negative coefficients, this variable is significant. The findings show that only two variables out of four are significant: GDP per capita from 1980 and average rate of investment. These variables have t-values of -2.540 and 3.319 respectively. Other variables have t-values that are lower then the critical value and hence are not significant.

The second regression analysis according to Granato et al. (1996) includes two social variables: achievement motivation index and postmaterialism index. In my case the F value of the regression is 5.740 (Table 3) which is rather high and this allows me to assume that relationship between dependent and two independent variables is significant.

On the other hand the power of independent variables to explain the variation of dependent variable is rather low since R² is 0.187 and adjusted R² is even lower 0.154 This finding can be explained by limited number of independent variables in this regression. Cook’s distance does not show any influential observations (maximum value at 0.652<1) (Table 3).

Concerning the t-test for the significance of each particular variable the critical value is 2.021, and both variables: achievement motivation index and 4-item postmaterialism index have proved to be significant with t-test values of 3.081 and -2.321 respectively (Appendix C).

In the next regression I include all of the independent variables from two regressions above. The F value of this regression is 3.575 (Table 3) which is higher then critical value 2.74, which shows that relationship between dependent variable and the set of

(33)

of adjusted R² is 0.229 (Table 3), Cook Distance does not show significant outliers (max value 0.345<1). The t-test shows three significant variables: GDP per capita 1980 (- 2.221), Investment (2.545) and achievement motivation index (2.054) (Appendix C).

Finally I conduct regression analysis that includes only significant variables form above.

In this regression the value of F is 6 and the values of R² and adjusted R² are 0.269 and 0.224 respectively. Cook Distance also does not show any abnormalities in the given data set (0.456<1) (Table 3). In this regression the significance of the three variables has increased, this can be seen form the values of the t-test: GDP per capita 1980 (-2.957), investment (2.774) and achievement motivation index (2.270) (Appendix C).

Regression Analysis with 12-item Postmaterialism Index

In addition to four regressions given above, I would like to perform another regression that includes 12-item index for postmaterialism, instead of 4 item index. The main reason for this is that 4-item index has received a substantial amount of criticism and I would like to check if the new 12-item index has a better explanatory power compared to the old 4-item index. Unfortunately the 12-item index is not available for all 53 countries that were used in the earlier regressions; therefore I have to limit the number of observations to 37 countries. Since the previous regressions have shown that only GDP1980, investment and achievement motivation index are significant variables when explain GDP growth per capita. I decided simply to add 12-item postmaterialism to these variables in order to see if this improved index will prove to be significant compared to the old 4-item index. The F-test in this regression shows that indeed the relationship between dependent and set of independent variables is significant, the F-value is 7.866 (Table 3), which is well above the critical value. The R² value and R² adjusted have increased compared to the prior model and amount to 0.496 and 0.433 respectively (Table 3). Finally all of the variables except GDP1980 are significant, nevertheless GDP1980 still has a negative relationship with GDP growth. T-values for Investment, Achievement and 12-item Postmaterialism Index are 2.521, 3.208 and -2.499 respectively (Appendix C).

(34)

Finally I would like to link the results to hypothesis. The findings show that three out of six hypotheses are supported. First hypothesis that predicted negative relationship between initial levels of GDP per capita and economic growth, was supported. Second hypothesis that predicted positive relationship between primary education and rate of economic growth, was not supported, moreover the calculated sign was opposite from the predicted one. Third hypothesis concerning the effect of secondary education on the economic growth of the country was not supported although it had a predicted sigh. The fourth hypothesis that predicted positive relationship between investment and economic growth was supported and it had a predicted sign. The fifth hypothesis concerning the positive relationship between achievement motivation and economic growth was also supported and had a predicted sign. The sixth and last hypothesis predicted negative relationship between postmaterialism and economic growth was supported and had a predicted sign.

Table 4. Hypotheses Summary

Hypothesis Expected Sign Real Sign Supported/Unsupported

H1 - - Supported

H2 + - Unsupported

H3 + + Unsupported

H4 + + Supported

H5 + + Supported

H6 - - Unsupported

5.4 Regression Analysis for Developed and Less Developed Groups of Countries

In addition to the findings above I would like to point out that economic growth in percentages is a very relative measure and does not mean the same rate of growth from country to country. In more developed countries absolute economic growth will have a smaller value when expressed in percentages, relative to the country with less developed

(35)

economy where high relative growth can be achieved with relatively low investments and improvements in economy. To deal with this problem all of the observed countries should be divided into two groups: developed and less developed. This technique will allow to identify the set of variables which is most responsible for growth, both in developed and developing economies.

Table 5a. OLS Estimation of Economic Growth Models. Dependent Variable: Mean Rate of Per Capita Economic Growth (1980-2000) – Less Developed Countries

Model Variable Model 1 Model 2 Model 3 Model 4 Constant -6.058E-02

(1.725)

9.226*

(3.180)

4.60 (4.272)

3.340*

(0.519) Per Capita GDP in 1980 -5.086E-04*

(0.000) -4.216E-04*

(0.000) -3.851E-04*

(0.000) Primary education 1980 -7326E-03

(0.021) -6.146E-03

(0.020) Secondary education 1980 1.964E-02

(0.019) 8.937E-03

(0.018) Investment .180*

0.074

0.105 (0.079)

Achievement Motivation 2.030*

(0.666) 1.979*

(0.760) 2.642*

(0.659) Postmaterialism 4-item -3.979*

(1.818) -1.566 (2.481)

R² 0.316 0.365 0.505 0.416

R² adjusted 0.202 0.318 0.370 0.372

SEE 1.819 1.678 1.617 1.611

F 2.776 7.762 3.738 9.597

Notes: Standard errors are in parentheses

* T-test significant at 95% confidence level.