Thesis BSc. Economics and Business
Economics and Finance
Faculty of Economics and Business
Macroeconomic of Happiness:
A Cross Country Analysis
By
Vania Esady
10003927
Supervisor: Matthias Weber
2
Abstract
In the past few decades, there is a growing demand for alteration in economic measurements. The recent availability of large cross-sectional happiness survey data has allowed economists to measure overall well-being. The purpose of macroeconomics of happiness study is to help government design better economic measurements that will enhance countries’ happiness. This thesis observes and compares how GDP per capita growth, inflation rate, unemployment rate and Gini index determine reported happiness in both the developed and non-developed countries. The paper concludes that even after controlling for personal characteristics, macroeconomic situations are important for individuals’ life satisfaction.
3
Table of Contents
Abstract
... 21. Introduction
... 42. Literature Review and Hypotheses Development
... 62.1 Easterlin Paradox and Income Hypothesis ... 6
2.2 Methodologies of Prior Studies on Happiness ... 7
2.3 Empirical Findings and Hypotheses ... 8
3. Research Methodology and Data
... 113.1 Research Methodology ... 11
3.2 World Values Survey ... 11
3.3 Macroeconomic Indicators ... 13
4. Results and Discussion
... 154.1 Descriptive Statistics ... 15
4.2 OLS Regressions ... 18
4.3 Test of Equality of Coefficients ... 25
5. Conclusion
... 274
1. Introduction
In April 2012, over six hundreds academic and political leaders from around the world gathered in the United Nations High Level Meeting for Wellbeing and Happiness: Defining a New Economic Paradigm. It depicts nations’ thirst for a transformation in economic measurement, to go from mere economic figures metric to overall well-being. The meeting was proposed by the Kingdom of Bhutan whom in 2013 has officially become the first country to imply Gross National Happiness. GNH measures the quality of life in a more holistic view unlike GDP, which often receives criticism for being a poor indicator of social welfare.
Should economists study happiness? The natural answer is yes, because the study of subjective well-being is correlated with observable phenomena and it is presumably important (Oswald, 1997). The term “happiness” was defined by Blanchflower (2007) to be the degree of which an individual judges the overall quality of his or her life. It is important to note that the term “happiness”, “life satisfaction” and “well-being" will be used interchangeably throughout the paper. The purpose of studying macroeconomics of happiness is to help government optimise the design for public policies which will enhance countries’ well-being. For the last few decades economists have been able to report empirical evidence on the correlation of happiness and macroeconomic events such as high economic growth, business cycle volatility and distribution of welfare. Macroeconomic of happiness is usually studied generally at international level or more often, focused on a particular event and selected countries. For example, Alesina et al. (2004) studied inequality and happiness in the U.S. and Europe, while Graham and Felton (2006) for Latin America. However, both studies did not embrace cross-country-groups comparative study.
The prospect of this thesis is to contribute on the cross countries comparative studies by examining the life satisfaction and macroeconomic indicators of 57 countries which are then sub-grouped into 21 developed and 36 non-developed countries. These indicators are expected to have distinctive impact on to life satisfaction in the country sub groups because the indicators are product of economic and political decisions (Di Tella et al., 2003). The expectation that these indicators determine happiness in different ways across countries is the ground of this research. The thesis looks out for which macroeconomic
5
indicators have successfully determined happiness or which still lacks of contribution in the developed and non-developed countries. At the end of this thesis, there is a hope to put forward new insights for policy making regarding overall well-being.
The central research question of this thesis is “How do the macroeconomic indicators
determine reported well-being in both, the developed and non-developed countries?” 1
In order to answer the above, this thesis evaluates how GDP per capita growth, inflation rate, unemployment rate and Gini index determine life satisfaction differently between the developed and non-developed countries. In this respect, a test of equality of the coefficients was done to provide certainty that the macroeconomic indicators explain happiness differently between the developed and non-developed countries. The life satisfaction data have been gathered through World Values Surveys. In this thesis, life satisfaction is estimated using OLS regression.
The regression results serve three findings. The first discovery of the thesis is that growth in GDP per capita has a significantly positive impact on life satisfaction. Then, inflation and unemployment is found to be costly not only to the unemployed but also to the entire country well-being. The last finding is surprising as inequality – based on Gini index – appears to positively determine happiness. The conjecture is, today’s inequality provides future opportunities.
This paper contributes to economics of happiness literature by closely examining and comparing the impact of particular macroeconomic indicators on happiness in different level of economies (i.e. develop countries and developing and transition economies). In a societal context, the findings offer insights to enable policymakers to improve policy which will maximise overall well-being.
The thesis is organised as follow. Section two provides Literatures Review and Hypothesis
Development and section three Research Methodology and Data. Section four presents Results and Discussion and section five gives Conclusion.
1 The decision whether a country belongs to developed or non-developed group is based on United Nation
definition, to be found in
http://www.un.org/en/development/desa/policy/wesp/wesp_current/2012country_class.pdf
6
2. Literature Review and Hypotheses Development
The literature on happiness is numerous and extensive. For the purpose of effective review, this thesis concentrates on literatures that have analysed the correlation between happiness and macroeconomic determinants. The first subsection of literature review will look at income-happiness (Easterlin) paradox and the relative income hypothesis; followed by examining methodologies on prior studies of happiness; and finally, the last subsection discusses profound empirical findings and the hypothesis development.
2.1 Easterlin Paradox and Income Hypothesis
In 1974, Richard Easterlin made a major contribution to the economics of happiness study through his empirical research. The Easterlin Paradox offered three claims regarding the correlation of income and happiness. The first claim is within a society there is a consistent finding of positive relationship between income and happiness. A more recent work of Easterlin (1995) reconfirmed that within a country at a given time those with higher incomes are, on average, happier. The second one claims that across countries, where cultural and social norms differ, the relationship is weaker and ambiguous. These first two claims are the result of cross-sectional studies. The last claim is that over time, greater income does not give rise to happiness. This was the result of a time series study for the United States which showed since 1941 the happiness level in the U.S. is almost flat.
By now, numerous scholars have attempted to explain the Easterlin Paradox through relative income hypothesis.2 The relativity can be explained in three perspectives: social
comparison, aspiration, and/or adaptation level. Easterlin (1974) explains relativity of happiness through social comparison which translates that individuals’ utility comes from a comparison of themselves with others close to them. Then, according to aspiration level view, happiness is determined by the gap between aspiration (expectation of the future) and achievement (success of the past) (Michalos 1991 and Inglehart 1990). Last but not least, the adaptation level view explains happiness as a function of individual’s income through time. In the short run, an increase in income
2 Oswald (1997) reviewed the work of Easterlin (1974), Hirsch (1976), Scitvosky (1976), and Frank (1985)
7
may increase happiness but in the long run, as the individual adapt to the new income, his utility will return to initial state (Diener et al., 1999).
2.2 Methodologies of Prior Studies on Happiness
The standard econometric approach to study economics of happiness is to estimate regressions using Ordinary Least Square (OLS), Ordered Probit or Ordered Logit. Generally, it makes little or no difference whichever estimation method to use, although, the size of the coefficients will be different (Blanchflower, 2007). Moreover, cross-sectional and panel countries are most often used when studying happiness. Only a limited number of studies have utilised time-series data.
One of the largest cross-national comparisons of happiness study was done by Inglehart et al. (2008) who used World Values Survey and European Values Survey to observe more than 50 countries between 1981 and 2007. Veenhoven (2014) covered a larger scope of study for 164 nations between 1945 and 2013. Both found that people in poorer countries have a much lower level of self-reported happiness. Also, well being is highly responsive to the satisfaction of basic needs but almost invariant to income at higher levels of development.
Di Tella et al. (2003) studied two international data on the reported well-being using panel-approach. The first random sample is taken from the Euro-Barometer Survey Series (1975-1992) where individuals were asked questions about happiness but as the word happiness translates differently in many languages, a question on life satisfaction is asked instead. More importantly, life satisfaction data was utilised in place of happiness because happiness and life satisfaction are correlated (0.56 for the period 1975-1986) and that life satisfaction data is available for a longer period of time – from 1975 to 1992. The second well-being data is from the United States General Social Survey (1972-1994) where a similar life satisfaction question was asked. They proceed with running ordered probit regressions for happiness on GDP per capita, inflation rate, unemployment rate and benefit replacement rate which were retrieved from various OECD data sets. They first looked at the effect of GDP per capita before continuing to build on the model by adding other macro variables and then controlling for a vector of personal variables as well as including a country fixed effect and a year fixed effect. They concluded that macroeconomic matters and there is a clear microeconomic pattern in the study of
8
happiness. The development of research methodology of this thesis will follow the steps taken by Di Tella et al. (2003). Finally, there are three main findings in the paper, those are, effect of GDP on happiness, cost of recessions, and happiness evidence on the role of the welfare state. An extended discussion of findings and results is provided in next subsection.
2.3 Empirical Findings and Hypotheses
First of all, there is a broad consensus upon the need of controlling for personal characteristics such as age, education, marital status and employment status. Di Tella et al. (2003) found a clear microeconomic pattern on the well-being levels of their data in a number of countries. Blanchflower and Oswald used World Values Survey (WVS) sample which shows a strong U-shaped pattern (the coefficient of age is found to be negative and age2 is positive). Helliwell (2003) attest education variable to be among the
weakest in WVS data. It is also seen as an imperfect guide because education ages and quality differ much from country to country. But Di Tella et al. (2003) found that those in higher education show higher happiness. Moreover, Helliwell also test for marital status and employment status. The result is married people are happiest, followed by living as married, widows or widowers, divorced and separated. Helliwell also confirmed the earlier findings by Clark and Oswald (1994) and Di Tella et al. (2001) that individuals report a large reductions in life satisfaction from being unemployed.
Moving ahead to the macroeconomic of happiness analysis, Di Tella et al. (2001) proved that there is strong correlation between people’s happiness and current and change-in-GDP per capita. Di Tella et al. (2003) found persuasive evidence for U.S. and Europe that well-being is robustly correlated with growth of GDP per capita, in which is consistent with the adaptation theories. This means that some of the well-being gains from extra national income will wear off over time. Moreover, Blanchflower (2007) examined in more detail how GDP per capita impacts the life satisfaction levels of the poorer European countries and found evidence that higher GDP per capita has a positive effect for these countries. This reasoning forms the basis for the first hypothesis.
9
Prior studies show that higher unemployment and inflation depresses subjective well-being (Blanchflower, 2007). Public appears to be extremely averse to unemployment. According to Di Tella et al. (2003), in order to keep their life satisfaction constant, individuals need to be compensated with approximately $200 per year. This is well above the loss of income due to recession. Along the line, Frey and Stutzer (2002) find that the drop of happiness due to unemployment is worse when it features psychological and societal effect, for instance, a loss of self-esteem and personal control, and, that being unemployed is heavily disgraced in some societies.
Di Tella et al. (2003) run a happiness regression on the level of benefits (to proxy for the cost of risk) and the unemployment rate to estimate at least a part of the consequences of being unemployed for an employed person. In principle, the estimation could allow policymakers to then compare the effects on happiness for workers of receiving their safety net with the loss from higher unemployment rates. The result is, individuals who live in Ireland are willing to pay up to US$214 to live in a country with more generous welfare state such as France. Moreover, Di Tella et al. also calculated the marginal rate of substitution between GDP per capita and inflation. Using the ratio of the two coefficients on GDP per capita and the inflation rate found in their regression result, an individual would have to be given compensation of approximately US$70 for each one percentage point rise in inflation. These findings form the basis for the second and third hypothesis.
Hypothesis 2: Assuming unemployment benefit is limited in non-developed countries, the cost of
unemployment on life satisfaction is expected to be more significant in non-developed than developed countries.
Hypothesis 3: In general, an increase in inflation rate lowers reported life satisfaction.
The last empirical finding to be reviewed is inequality. Inequality discussed in economics of happiness studies are based on either reported individual income inequality from surveys, or a quantitative measure of inequality such as Gini index, Hoover index and Palma ratio. Alesina et al. (2004) found a clear negative effect of the Gini coefficient on satisfaction for the USA and various European countries. According to Alesina et al. (2004), there are two broad reasons why inequality may affect individual well-being. The first is because individuals may have a taste for equality and the alternative is that people
10
regard income inequality as an indicator of their future income. More specifically, the income-equality-as-taste hypothesis predicts that the negative effect of inequality on happiness should be stronger amongst the rich. On the other hand, the hypothesis of inequality-as-predictor-of-future-income predicts that the effect particularly depend quite crucially on the degree of mobility that is present in society. Based on these predictions, the last hypothesis of this paper is:
11
3. Research Methodology and Data
3.1 Research Methodology
The objective of the research is to investigate the relationship between Life Satisfaction and GDP per capita growth, inflation rate, unemployment rate, and Gini index while controlling for a vector of personal variables: age, marital status, employment status and highest education level attained. The estimation method used throughout this thesis is Ordinary Least Squares (OLS). OLS is used in order to keep simplicity, following the method of other authors (Agan et al. 2009, Blanchflower 2007 and Inglehart et al. 2008). The first regression of life satisfaction is on GDP per capita (taken in log), unemployment, and inflation, in which are the three most observed variables in economics of happiness studies. The second specification adds on Gini index to account for inequality. The third regression will then control for personal variables. These specifications are regressed separately for all countries, developed countries and non-developed countries samples. In formal term, the model specification for this thesis is:
Where Life Satisfactionic represents reported life satisfaction score of individual i in
country c. Log(GDP per cap)c is GDP per capita growth in country c; Inflationc is the inflation rate of country c; Unemploymentc is the unemployment rate of country c, and
Ginic is the Gini index of country c. Ageic, Marital static, Highest eduic, and Employ static, represents the age, marital status, highest level of education attained, and employment status of individual i in country c, respectively.
3.2 World Values Survey
Literature on economics of happiness relies heavily on surveys of the reported well-being (Graham, 2005). The dependent variable, life satisfaction, and a vector of personal variables of this thesis are retrieved from the integrated dataset of World Values Survey
12
(WVS). Echoing the decision of Inglehart et al. (2008), this thesis utilizes World Values Survey dataset because it is one of the best sources available today for international comparisons of life satisfaction. WVS has been conducted in almost 100 countries which cover about 90% of the world’s population. These surveys have been implemented in six different waves from 1981 to 2014. This paper will focus on Wave 5 (2005-2009) because it has the most number of countries (58) and observations (83,975) within a wave. There are over 260 variables reported in Wave 5 but to fit the purpose of this paper, only five variables will be used, those are, Life Satisfaction, Age, Marital Status, Highest Level of Education Attained, and Employment Status.
The variable of reported happiness (named V10) is measured in 4 categories: very happy, quite happy, not very happy, and not at all happy. However, instead of happiness, the variable of reported life satisfaction (named V22) is used in this thesis because life satisfaction has a broader range of scale, from 1 (most dissatisfied) to 10 (most satisfied). Subsequently, as written by Wolfers (2003), although happiness and life satisfaction concept may differ, responses are highly correlated and joined under “subjective well-being”. The question “All things considered, how satisfied are you with your life as a whole
these days?” was asked throughout all waves to measure life satisfaction. The results of
‘missing or not asked by the interviewer’, ‘no answer’ and ‘don’t know’ are omitted from analysis.
As the paper focuses on the difference of life satisfaction between developed and non-developed countries, it is worth to look at the average life satisfaction scores. Table 1 shows the average life satisfaction scores of developed countries, while Table 2 shows the scores of non-developed countries, both in ranking order in each category. Switzerland has the highest score (8.056) and Bulgaria has the lowest score (4.944) among developed countries. Among non-developed countries, Columbia scores highest (8.286) and Iraq scores lowest (4.397).
The reliability of data usage of Wave 5 may be questioned because of the financial crisis that started in 2008. Although not included in this paper, the average life satisfaction scores of Wave 5 does not demonstrate exceptional differences with that of other waves. With this reason and also the fact that the surveys were done largely and separately under various times, data analysis will not be affected by a possible existence of financial crisis.
13
Source: WVS 5: 2005 to 2009
Source: WVS 5: 2005 to 2009
3.3 Macroeconomic Indicators
The data of macroeconomic indicators to be analysed in this thesis, including GDP per capita, inflation rate, unemployment rate, and Gini index are taken from The World Bank national accounts data, International Labour Organization (ILO), and OECD National Accounts data files. Data of Taiwan is not available in these sources and
Table 1 – Average Life Satisfaction in Developed Countries
Ranking Country Life Satisfaction Ranking Country Life Satisfaction
1 Switzerland 8.056 12 Australia 7.212 2 Norway 7.936 13 Slovenia 7.211 3 Finland 7.836 14 Germany 7.057 4 Netherlands 7.744 15 Poland 6.931 5 Sweden 7.735 16 France 6.901 6 Canada 7.716 17 Japan 6.859
7 New Zealand 7.612 18 Italy 6.837
8 Great Britain 7.572 19 Hungary 5.875
9 Cyprus 7.372 20 Romania 5.243
10 Spain 7.285 21 Bulgaria 4.944
11 USA 7.245
Table 2 - Average Life Satisfaction in Non Developed Countries
Ranking Country Life Satisfaction Ranking Country Life Satisfaction
1 Columbia 8.286 20 Indonesia 6.414
2 Mexico 7.924 21 Hong Kong 6.355
3 Guatemala 7.916 22 South Korea 6.338
4 Argentina 7.727 23 Ghana 6.089 5 Brazil 7.620 24 Russia 6.025 6 Turkey 7.459 25 Zambia 5.880 7 Uruguay 7.392 26 Serbia 5.747 8 Trinidad 7.300 27 Egypt 5.737 9 Thailand 7.201 28 Ukraine 5.634 10 Andorra 7.132 29 India 5.630 11 Jordan 7.086 30 Mali 5.547 12 Chile 7.086 31 Moldova 5.420 13 Vietnam 7.016 32 Burkina 5.405
14 South Africa 6.989 33 Morocco 5.234
15 Peru 6.974 34 Rwanda 4.949
16 Malaysia 6.830 35 Ethiopia 4.948
17 Taiwan 6.584 36 Georgia 4.857
18 China 6.581 37 Iraq 4.397
14
therefore it has been omitted from the analysis. There are now 57 countries (21 developed countries and 36 non-developed countries) available for this research. Detailed descriptions of the macroeconomic indicators are provided below.
In this thesis, the variable named “log (GDP per capita)” represents the GDP per capita growth. GDP per capita (current US$) itself is “the sum of gross value added by all resident producers in the economy plus any product taxes and minus any subsidies not included in the value of the products” 3 divided by midyear population.
Furthermore, the two ways to measure inflation are through Consumer Price Index (CPI) and GDP deflator. Inflation CPI “reflects the annual percentage change in the cost to the average consumer of acquiring a basket of goods and services” and GDP deflator is the ratio of GDP in current local currency to GDP in constant local currency. 4,5
Inflation CPI is used in the specifications to fit our purpose of studying individual life satisfaction.
Additionally, there are various categories of unemployment data available in the World Bank dataset. This thesis uses the total unemployment (% of total labour force), a modelled ILO estimate to keep simplicity and be in line with the works of others. In this modelled ILO estimate, unemployment refers to the share of the labour force that is without work but available for and seeking employment.6
Finally, Gini index measures the extent to which the distribution of income or consumption expenditure among individuals or households within an economy deviates from a perfectly equal distribution. Gini index of 0 represents perfect equality, while an index of 100 implies perfect inequality. 7 Due to many missing Gini index from the
World Bank dataset, the Gini index of developed countries is then gathered from the OECD database which provide similar index numbers.
3 Source: http://data.worldbank.org/indicator/NY.GDP.PCAP.CD, dated 1 June 2014 4 Source: http://data.worldbank.org/indicator/FP.CPI.TOTL.ZG, dated 1 June 2014 5 Source: http://data.worldbank.org/indicator/NY.GDP.DEFL.KD.ZG, dated 1 June 2014 6 Source: http://data.worldbank.org/indicator/SL.UEM.TOTL.ZS, dated 1 June 2014 7 Source: http://data.worldbank.org/indicator/SI.POV.GINI and
15
4. Results and Discussion
4.1 Descriptive Statistics
Table 3 summarises the descriptive statistics of variables used in this research for the full data set, that is, all countries. The average (median) score of Life Satisfaction is around 6.606 (7.000). The averages (medians) of the macroeconomics variables in 57 countries are around: 8.71% (8.61%) for GDP per capita growth, 6.73% (5.21%) for inflation rate, unemployment rate for 8.09% (7.02%) and 39.00% (35.66%) for Gini index. The average (median) age is around 41.32 (39) years old.
Table 4 and 5 summarises the descriptive statistics of variables of the segregated country groups, developed and non-developed. The average (median) score of Life Satisfaction in develop countries is around 7.052 (8.000) and in non-developed countries is 6.412 (7.000). The life satisfaction score is shown to be higher in developed country. Diener et al. (1995) and Inglehart (1990) provided similar evidence in their studies that, on average, persons living in rich countries are happier than those living in poor countries.
The averages (medians) of the macroeconomics variables in the developed countries are around: 10.30% (10.56%) for GDP per capita growth, 2.68% (2.49%) for inflation rate, unemployment rate for 6.58% (6.68%) and 30.68% (31.39%) for Gini index. The average (median) age is around 47.44 (47) years old. On the other hand, The averages (medians) of the macroeconomics variables in the non-developed countries are around: 8.01% (8.06%) for GDP per capita growth, 8.63% (7.20%) for inflation rate, unemployment rate for 6.58% (6.68%) and 42.84% (40.92%) for Gini index. The average (median) age is around 38.67 (36) years old. The average GDP per capita growth is higher in developed countries while the average of inflation rate, unemployment rate and Gini index is lower in developed than in non-developed countries. The median age is older in developed countries.
Table 6 shows the Pearson (top) – Spearman (bottom) correlation matrix based on full sample dataset. GDP per capita growth, Gini index, and highest level of education attained are positively correlated with life satisfaction. On the other hand, inflation, unemployment, age, marital status and employment status are shown to have negative correlation with life satisfaction.
16
Table 3: Desriptive Statistics - All countries
Variable Mean Std. Err Median Mode Std. Dev Variance Kurtosis Skewness Min. Max. Count 99th% 1st%
Life Satisfaction 6.606 0.009 7.000 15.000 2.490 6.199 3.510 -0.793 -5.000 10.000 83975 10.000 -1.000 Log (GDP per cap) 8.714 0.005 8.616 5.775 1.438 2.068 2.168 -0.114 5.505 11.280 82748 11.280 5.505 Inflation 6.738 0.018 5.218 19.921 5.067 25.670 3.241 1.018 0.010 19.932 79743 19.932 0.010 Unemployment 8.096 0.017 7.020 22.420 4.890 23.910 4.325 1.172 0.600 23.020 81745 23.020 0.600 Gini 39.001 0.037 35.660 41.016 10.215 104.354 2.976 0.921 24.254 65.270 77241 65.270 24.254 Age 41.324 0.057 39.000 103.000 16.649 277.201 2.483 0.466 -5.000 98.000 83975 81.000 17.000 Marital stat 2.703 0.008 1.000 11.000 2.193 4.811 1.668 0.629 -5.000 6.000 83975 6.000 1.000 Highest edu 5.193 0.009 5.000 14.000 2.555 6.530 2.213 -0.161 -5.000 9.000 83975 9.000 1.000 Employ stat 3.226 0.008 3.000 13.000 2.425 5.883 2.819 -0.106 -5.000 8.000 83975 8.000 -4.000
Source: WVS 5, the World Bank, ILO, and OECD: 2005 to 2009
Table 4: Desriptive Statistics - Developed Countries
Variable Mean Std. Err Median Mode Std. Dev Variance Kurtosis Skewness Min. Max. Count 99th% 1st%
Life Satisfaction 7.052 0.014 8.000 15.000 2.199 4.837 5.404 -1.292 -5.000 10.000 25406 10.000 -1.000 Log (GDP per cap) 10.296 0.004 10.564 2.683 0.670 0.449 3.559 -1.238 8.597 11.280 25406 11.280 8.597 Inflation 2.684 0.011 2.494 7.151 1.741 3.031 4.321 1.409 0.010 7.161 25406 7.161 0.010 Unemployment 6.585 0.014 6.680 8.020 2.167 4.695 2.637 0.498 3.260 11.280 25406 11.280 3.260 Gini 30.677 0.021 31.386 13.686 3.221 10.372 2.924 -0.046 24.254 37.941 24356 37.941 24.254 Age 47.437 0.110 47.000 100.000 17.462 304.909 2.206 0.117 -2.000 98.000 25406 84.000 18.000 Marital stat 2.557 0.013 1.000 11.000 2.095 4.389 2.025 0.702 -5.000 6.000 25406 6.000 1.000 Highest edu 5.779 0.015 6.000 14.000 2.347 5.509 3.020 -0.418 -5.000 9.000 25406 9.000 1.000 Employ stat 3.025 0.013 3.000 13.000 2.099 4.406 2.483 0.474 -5.000 8.000 25406 8.000 1.000
17
Table 5: Desriptive Statistics - Non Developed countries
Variable Mean Std. Err Median Mode Std. Dev Variance Kurtosis Skewness Min. Max. Count 99th% 1st%
Life Satisfaction 6.412 0.011 7.000 15.000 2.582 6.666 3.083 -0.613 -5.000 10.000 58569 10.000 -1.000 Log (GDP per cap) 8.013 0.005 8.064 5.041 1.089 1.185 2.910 -0.144 5.505 10.546 57342 10.546 5.505 Inflation 8.634 0.021 7.198 17.964 4.998 24.979 2.675 0.725 1.968 19.932 54337 19.932 1.968 Unemployment 8.777 0.023 8.240 22.420 5.575 31.082 3.173 0.830 0.600 23.020 56339 23.020 0.600 Gini 42.835 0.044 40.920 36.996 10.051 101.030 2.413 0.592 28.274 65.270 52885 65.270 28.274 Age 38.672 0.064 36.000 102.000 15.555 241.943 2.754 0.593 -5.000 97.000 58569 78.000 16.000 Marital stat 2.767 0.009 1.000 11.000 2.232 4.981 1.532 0.595 -5.000 6.000 58569 6.000 1.000 Highest edu 4.938 0.011 5.000 14.000 2.600 6.760 2.031 -0.036 -5.000 9.000 58569 9.000 1.000 Employ stat 3.313 0.011 3.000 13.000 2.549 6.498 2.874 -0.279 -5.000 8.000 58569 8.000 -4.000
Source: WVS 5, the World Bank, ILO, and OECD: 2005 to 2009
Table 6: Pearson-Spearman Correlation Matrix
Life
Satisfaction Log (GDP per cap) Inflation Unemployment Gini Age Marital stat Highest edu Employ Stat
Life Satisfaction 0.266 -0.2401 -0.0489 0.0838 -0.0057 -0.011 0.1601 -0.0856
Log (GDP per cap) 0.2529 -0.6613 -0.0055 -0.3323 0.2381 0.0024 0.2453 -0.1173
Inflation -0.2478 -0.6012 0.2875 0.1789 -0.2168 0.0133 -0.0867 0.1377 Unemployment -0.0527 -0.0315 0.3114 0.0444 -0.0558 0.0487 0.0515 0.0583 Gini 0.0953 -0.307 0.0147 0.2995 -0.1756 0.1003 -0.0887 0.039 Age -0.0075 0.2605 -0.2085 -0.0657 -0.1668 -0.3956 -0.1621 -0.0838 Marital stat -0.0203 -0.0325 0.0374 0.0613 0.1152 -0.3146 0.1184 0.1391 Highest edu 0.1632 0.2509 -0.0825 0.0439 -0.0676 -0.1607 0.1043 -0.1641 Employ stat -0.0804 -0.0967 0.1357 0.0566 0.0495 -0.0403 0.1521 -0.1358
18
4.2 OLS Regressions
In this section, OLS regression results for all countries, developed countries and non-developed countries are presented in Table 7. It is important to note that correlation between life satisfaction and macroeconomics and personal variables is implicitly assumed to imply causality.8 The causality is also presumed to run from regressors to life
satisfaction and not the other way around.9
The first apparent finding is the significant personal pattern found in the regression results. All personal variables are significant at 0.001 significance level. “Highest
Education Level Attained” coefficients are displayed to have largest magnitudes than other
personal variable throughout all related regression models. From this result, it can perhaps be deduced that education largely determines happiness and that a higher education positively affect individuals’ life satisfaction. Unfortunately, the estimation method used in this thesis is too limited to make precise claims. A study that uses Ordered Probit or Ordered Logit as estimation method may be able to estimate further, for example, “Which level of education (high school, college, undergraduate, etc.) can most positively affect life satisfaction?” Also, since Age, Marital Status, and Employment
Status have sub categories within them, it is also not possible to accurately calculate how
they determine happiness. Nevertheless, after controlling for some personal characteristics, macroeconomic indicators are shown to have statistically robust effects on reported life satisfaction.
Another aspect to consider is the R2 and adjusted R2. The magnitudes of both R2 and
adjusted R2 increase as additional variables are added to the models which reflect
improvement of the models in general. However, the sizes are still considerably small. This is a regular finding when working with models that predict human behaviour. Ultimately, below are discussions of the regression results. The findings are put together into three sections: Effect of a GDP per capita growth on Life Satisfaction; the Costs of Inflation
and Unemployment and last but not least, Happiness Evidence on the Role of Equality
8 Frey and Stutzer (2002) provides for a more elaborative discussion
9 Easterlin (1974) was inclined to interpret the causality as running from income to happiness. Also, Frey
and Stutzer (2002) concluded from studies of Winkelmann and Winkelmann (1998) and Marks and Fleming (1999) that the main causality seems clearly to run from unemployment to unhappiness.
19
Table 7: Life Satisfaction Equation, OLS: 2005 to 2009
Variable All Countries Developed countries Non-Developed Countries
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Log(GDP per cap) 0.273*** 0.492*** 0.455*** 0.747*** 0.789*** 0.814*** 0.405*** 0.568*** 0.505*** (-0.008) (-0.009) (-0.009) (-0.044) (-0.045) (-0.044) (-0.012) (-0.014) (-0.014) Inflation -0.071*** -0.026*** -0.032*** -0.136*** -0.123*** -0.111*** -0.076*** -0.032*** -0.035*** (-0.002) (-0.002) (-0.002) (-0.016) (-0.016) (-0.016) (-0.002) (-0.003) (-0.003) Unemployment 0.002 -0.046*** -0.046*** -0.012 -0.001 0.028*** -0.004 -0.049*** -0.051*** (-0.002) (-0.002) (-0.002) (-0.007) (-0.007) (-0.007) (-0.002) (-0.002) (-0.002) Gini 0.052*** 0.052*** -0.009* -0.011* 0.045*** 0.047*** (-0.001) (-0.001) (-0.004) (-0.004) (-0.001) (-0.001) Age -0.009*** -0.006*** -0.009*** (-0.001) (-0.001) (-0.001) Marital stat -0.061*** -0.092*** -0.047*** (-0.004) (-0.007) (-0.005) Highest edu 0.100*** 0.092*** 0.108*** (-0.004) (-0.006) (-0.004) Employ stat -0.033*** -0.060*** -0.024*** (-0.004) (-0.007) (-0.004) Intercept 4.671*** 0.870*** 1.398*** -0.202 -0.487 -0.718 3.835*** 0.719*** 1.220*** (-0.075) (-0.105) (-0.107) (-0.507) (-0.544) (-0.540) (0.097) (0.120) (-0.123) R-sqr 0.0728 0.1093 0.1132 0.1083 0.1127 0.1391 0.0622 0.1017 0.1178 Adj R-sqr 0.0727 0.1093 0.1132 0.1082 0.1125 0.1388 0.0621 0.1017 0.1176 RMSE 2.442 2.387 2.376 2.077 2.078 2.047 2.530 2.503 2.480 No. of Obs 79743 75239 75239 25406 24356 24356 54337 50883 50883
20
Effect of a GDP per capita Growth on Life Satisfaction
Resounding the findings of Easterlin (1974) and Di Tella et al. (2003), there is enough evidence in our data that GDP per capita growth produces higher happiness. In Table 7 the coefficients of log (GDP per capita) are positively significant and also appear to be the largest determinant of life satisfaction compared to other macroeconomic variables. Next, the result of all countries sample will be used to predict the effect of a GDP per capita growth on life satisfaction in units of thousands of dollars.
In Table 7, for all countries sample, the sizes of the coefficients grew larger when Gini index and personal variables are added to the regressions, that is 0.272, 0.492, and 0.454 for column (1), (2), and (3) respectively. In column (3), a one-percentage-point change in log (GDP per capita) is associated with a change in life satisfaction of 0.00454 (computed by 0.01*0.454). Furthermore, these coefficients can be used to estimate the effect of growth in GDP per capita on life satisfaction, in units of thousands of dollars. For example, in column (1) of Table 3, an increase of average income from $10,000 to $11,000 will have an estimated value change in life satisfaction of 0.0259 (the number comes from 0.272*[ln (11) – ln (10)]). And when average income increases from $50,000 to $51,000, the estimated value of change is only 0.00539. This predicts that a $1000 increase in income has a larger effect on life satisfaction in poorer societies (non-developed countries) than it does wealthier societies ((non-developed countries).10 This is
usually known as “diminishing marginal utility of income”. From the data set used in this thesis, Table 3, 4 and 5 shows the average GDP per capita for all countries is $14,454, for developed countries $35,117 and for non-developed countries $5,299. If the prediction is true, then policy makers in non-developed countries can set target for higher GDP per capita growth in order to elevate the life satisfaction in their countries.
However, the prediction is not supported by the regression results presented in Table 7. The sizes of coefficients are much larger in developed than in non-developed countries. This suggests that an increase in income is relatively more influential in developed countries.
The coefficients of log (GDP per capita) in developed countries are 0.747, 0.789, and 0.814 for model 1, 2 and 3 respectively. While the coefficients of log (GDP per capita) in
21
non-developed countries are 0.405, 0.568 and 0.505 for model 1, 2 and 3 respectively. The relatively more significant correlation found in developed countries could be argued that individuals living in countries with higher income per-capita growth have more chances to boost their life satisfaction. However, this line of argument should be seriously considered in the future. Nonetheless, the result shows that a growth in GDP produces more life satisfaction.
The Costs of Inflation and Unemployment
This section has three objectives which focus on the cost of inflation and unemployment on well-being. The first is to show that individuals care about inflation and unemployment. The second is to measure the shortfall in life satisfaction due to unemployment. The last aim is to calculate the costs of inflation in terms of unemployment.
The findings present evidence that inflation and unemployment belong in a well-being function. Looking at all countries sample, inflation coefficients are negative and statistically significant at 0.001. This significance suggests that in general, inflation is disliked. An increase of annual rate of inflation will in fact lower life satisfaction score. This result is robust because as personal variables are inserted into the equation, the result coefficients of inflation remain significant. The result is similar when separated analysis is done on developed and non-developed countries. These findings indicate that inflation rate is indeed a crucial determinant of life-satisfaction. On the other hand, oddly enough, unemployment is statistically insignificant when the specification only include log (GDP per capita), inflation, and unemployment. It is also still insignificant when Gini enters the second life satisfaction model for developed countries. Other than that, unemployment has significant negative impact onto life satisfaction, especially when personal characteristics are controlled. Inflation and unemployment rate are not mere percentages but are implied policies of the central governments of each country. Its significance in determining well-being entails the need of careful policy making.
Moving onwards, the second aim concentrates on calculating the cost of increase of an increase in unemployment rate on life satisfaction. The second regression model of this thesis studies the dependency of life satisfaction on only macroeconomic variables while the third regression adds personal variables, including employment status. The inclusion
22
of employment status enables controlling for the personal cost of joblessness and to test for any indirect psychic losses from an increase of unemployment rate. In other words, an increase in unemployment can affect life satisfaction through at least two channels. One is the direct effect to individuals who become unhappy because they are unlucky enough to be unemployed. The other effect may enter indirectly through fear of being unemployed in the future and also an aversion of higher tax for unemployment contribution. In order to calculate the full losses, these two effects have to be added together.11 For instance, from column (2) in Table 7, it can be predicted that a one per
cent increase in the unemployment rate will lower life satisfaction score by approximately 0.046 for the average citizen. This number can be viewed as capturing the indirect effect mentioned above. And column (3) shows being unemployed create a loss of also 0.046. Therefore, the entire well-being cost of a one percentage point increase in the unemployment rate is given by the sum of the two components, that is, 0.092.
An interesting result is born when analysis for developed and non developed countries is done separately. A one per cent increase in unemployment will lower the entire well-being by 0.298 and 0.100 for developed and non-developed countries, respectively. Out of these summations, the personal effect of unemployment in developed country is only 0.001 (as found in column (5) of Table 7), a much smaller number than in non-developed countries, 0.049 (column (8) of Table 7). These numbers show that the personal cost of unemployment on overall well-being is fairly high. With respect to this, there is a need for revising and improving the social insurance and benefit policy, particularly in non-developed countries.
Another macroeconomic indicator that could depress the economies’ well being is inflation. Survey evidence presented by Shiller (1997) shows that in addition to regular loss of unemployment, individuals report a number of unconventional costs of inflation, such as exploitation, national prestige, and loss of morale. It is worth to note that in developed countries, inflation is the second largest macroeconomic determinant of life satisfaction after GDP per capita. While, as seen in column (8) and (9) of Table 7, in non-developed countries inflation rate has the least contribution to life satisfaction
11 The calculation method used here is as found in the paper of Di Tella et al. (2001). Throughout their
paper, unemployment and inflation are measured in fractions. For example, an 8-percent rate of inflation is entered in their data set as 0.08. When calculating the impact of a 1% increase in inflation, they
multiplied the coefficients with 0.01. But, since this thesis had not changed the measurement into fractions, we do not to multiply our coefficients with the corresponding percentage increase.
23
among other macroeconomic determinants. The differences can be explained by contemplating on the different monetary policy targets declared by central banks. While most of the non-developed countries listed in this thesis do not have inflation targeting, many of the developed countries set inflation targeting as their primary objective, for instance the European Central Bank and Central Bank of Canada aims for a less than or around 2% inflation rate. 12
The calculation of costs of inflation has been seen as a perplexed issue by economists (Mankiw, 1997). In this thesis, the cost of inflation is calculated in terms of unemployment, or often referred to as marginal rate of substitution between inflation and unemployment (Di Tella et al., 2001). In other words, in column (2) of Table 7, a one per cent increase in unemployment is equal to the loss by an extra 3.487 point of inflation. The 3.487 marginal rate of substitution comes from 0.091/0.0261, where 0.091 is the marginal effect of unemployment on well-being calculated earlier and 0.0261 is the marginal effect of inflation on well-being. In developed countries, the marginal rate of substitution is 2.423 (computed by 0.298/0.123) and in non-developed countries 3.125 (computed by 0.100/0.032). Thus, with these results provide convincing evidence that inflation is strong disliked by individuals and further, reducing inflation is costly in terms of extra unemployment.
Happiness Evidence on the Role of Equality
Understanding how inequality affects happiness can contribute to our comprehension of the political determinants of particular economic policy choices or support for redistribution (Graham and Felton, 2006). Many discussions make implicit assumptions and also show empirical results that individuals are inequality-averse (among many are, Di Tella et al. 2003, Blanchflower and Oswald 2004 and Alesina et al. 2004). However, it is also not rare to find studies which conclude that happiness and inequality are uncorrelated. Easterlin (1974), Helliwell (2003), and Ferrer-i-Carbonell and Ramos (2009) find that additional inequality variable provide insignificant explanatory power to the well-being equation.
12 According to IMF: https://www.imf.org/external/pubs/ft/fandd/2010/03/pdf/roger.pdf, dated 15
24
This thesis uses data of World Values Survey and Gini index from 2005-2009 to test the inequality-averse hypothesis. The result shows that inequality is unambiguously positively correlated with life satisfaction for the full sample, that is, all countries dataset. This result is similar with the finding of Ball (2001) and Clark (2003) which runs counter to the supposed public dislike of inequality.
To this extent, our result shows individuals appear to be inequality-loving rather than inequality averse. However, when the analysis is broke down into country sub-groups, the finding is not consistent. Indeed, it echoes Alesina et al.’s conclusion that the size of inequality effect depends on the functional forms being used, and may well differ by country. For non-developed countries, Gini are positive at 0.001 levels of significant while in contrary, Gini of developed countries are negative and less significant (significant at 0.05). More specifically, the magnitudes of the coefficients are also relatively larger in non-developed countries. For instance, a ten percentages point rise of inequality will increase the life satisfaction of non developed countries by 0.045 points and yet only lower life satisfaction in developed countries by 0.009 points. The considerably different magnitudes can be predicted by how extensive inequality matters to each group. Although comparatively fewer studies of inequality and happiness have been conducted in non-developed countries, the findings thus far show that inequality matters more in developing and transition countries (Grosfeld and Senik 2008 and Graham and Sukhtankar 2001) than it does in developed countries (Di Tella et. al., 2003) in which reveals modest effect or inconclusive evidence that inequality matters at all.
At last, discussions on economics equality suggest that inequality can either be a signal of income mobility and opportunity or a signal of injustice. The result of this thesis predicts inequality to be a signal of persistent advantage and future opportunities for the poor and persistent disadvantage for the wealthy, possibly because of an opposition for redistribution. The relationship between inequality and subjective well-being will continue to be an abundant ground for future research in social science.
25
4.3 Test of Equality of Coefficients
The macroeconomic indicators utilised in this thesis need to be proven that they are certainly different when determining happiness of developed and non-developed countries to ensure the need of separated countries groups’ analysis. In order to investigate this issue, test of equality of coefficients will be run. The focus of this test is put on the significance of dummy interacted variables coefficients. For the purpose of the test, dummy variable is generated to separate non-developed countries from developed countries (d=1 for developed and d=0 for non-developed).
The test proceeds by generating dummy interacted variables, presented in Table 8 in italics (e.g. d*inflation). Then, life satisfaction is regressed on various variables, dummy, and dummy interacted variables. For instance, the first regression will be life satisfaction on GDP per capita, inflation, unemployment, dummy, d*gdp per capita, d*inflation, and
d*unemployment. Notice that in Table 8 the coefficients of the log (GDP per capita),
inflation, unemployment and Gini and personal variables are the same with that in the regressions of non-developed countries in Table 7. Also, the sum of the variable coefficients of non-developed countries and coefficients of dummy interacted variables is equal to the variable coefficients of developed countries.
Turning to the results presented in Table 8, coefficients of d*log (GDP per capita) are significant at the level of 0.001, indicating that growth in income notably determine life satisfaction differently across developed and non-developed countries. The magnitudes of the coefficients are also remarkably different. The explanation of the major difference had been explained in OLS Regressions section.
Subsequently, the coefficients of d*Inflation in column (1) of Table 8 are significant at 0.05 and at 0.001 in column (2) to (3), respectively. This result points out that inflation play different role in determining happiness for developed and non-developed countries. In developed countries, rising inflation is presumed to be plainly unfavourable and therefore the central governments work hard to keep it at a low and stable rate. On the other hand, inflation in developing countries may still be linked with positive economic growth, which was discussed above, does increase happiness.
26
Table 8: Test of Equality of Coefficients
Variable (1) (2) (3)
Log(GDP per cap) 0.405*** 0.568*** 0.505***
(0.0109) (0.01295) (0.0132) Inflation -0.0760*** -0.0318*** -0.0350*** (0.00232) (0.00266) (0.00267) Unemployment -0.00368 -0.0493 -0.0511*** (0.00198) (0.00219) (0.00217) Gini 0.0452 0.0466*** (0.00127) (0.00128) Age -0.00892*** (0.000746) Marital stat -0.0474*** (0.00511) Highest edu 0.100*** (0.00362) Employ stat -0.0334*** (0.00363)
d*log(GDP per cap) 0.343*** 0.221*** 0.309***
(0.0513) (0.0525) (0.0521) d*Inflation -0.0598* -0.0909*** -0.0758*** (0.0183) (0.0185) (0.0183) d*Unemployment -0.00814 0.0479*** 0.0795*** (0.00814) (0.00842) (0.00848) d*Gini -0.0537*** -0.0571*** (0.00492) (0.00488) d*Age 0.00248* (0.00119) d*Marital stat -0.04456*** (0.00911) d*Highest edu -0.0163* (0.00809) d*Employ stat -0.0366*** (0.00853) Dummy -4.037*** -1.206 -1.938** (0.5916) (0.631) (0.630) Constant 3.835*** 0.719*** 1.22*** (0.0919) (0.114) (0.117) R-sqr 0.0891 0.1195 0.1378 Adj R-sqr 0.0890 01194 0.1376 RMSE 2.947 2.3735 2.3488 No. of Observations 79743 75239 75239
Notes: Standard errors are in parentheses. Significance is shown by asterisks
27
Unexpectedly, the coefficient of d*Unemployment in column (1) of Table 8 is insignificant. The explanation may lay on the fact that in column (1), Gini index was not yet incorporated which means the inequality was not yet considered. If supposedly equality is interpreted as equal income, thus in times of rising unemployment, equality would only be achieved by replacing income with a social insurance such as unemployment benefit. The presumption is perhaps true because when Gini is inserted, the coefficients of d*Unemployment become statistically significant in column (2) and (3) in Table 8. The significant coefficients also depict that life satisfaction in developed and non-developed countries is determined differently by unemployment.
Last but not least, the coefficients of d*Gini are significant at 0.001. It suggests that inequality - the gap between the poor and the rich - is in fact a very important macroeconomic indicator in determining happiness. Considering the income hypothesis, significant effect of inequality reflects the proposition that happiness is relative, especially by social comparison.
5. Conclusion
The thesis examines and compares how GDP per capita, inflation rate, unemployment rate and Gini index determine life satisfaction in both the developed and non-developed countries. The findings show that even after controlling for a vector of personal characteristics, the macroeconomic situations of a country are indeed important for determining individual life satisfaction. The empirical results show that economic boom which generates GDP per capita growth indeed increases life satisfaction, especially in developed countries. While an increase in annual unemployment and inflation rate decreases happiness. Approximately, half of the cost of unemployment on the entire sample population comes from personal effect which calls on upon the attention of policy makers for an improvement in the social insurance system. Then, the result shows inflation is generally disliked, more specifically, inflation largely depresses the life satisfaction of developed countries. This reflects why inflation targeting has been the primary goal of many central banks in developed countries for many years. Last but not least, surprisingly, the result shows that individuals appear to be inequality-loving rather than the usual prediction of being inequality averse.
28
Reference List
Agan, Y., Sevinc, E., and Orhan, M. (2009). Impact of main macroeconomic indicators on happiness. European Journal of Economic and Political Studies, 2, 13-21.
Alesina, A., Di Tella, R., and MacCulloh, R. (2004). Inequality and happiness: are Europeans and Americans different?. Journal of Public Economics, 88, 2009-2004. Ball, R. (2001). Incomes, inequality and happiness: new evidence. Haverford College,
mimeo.
Blanchard, O. J., & Fischer, S. (1989). Lectures on macroeconomics. MIT press.
Blanchflower, D. G. (1991). Fear, unemployment and pay flexibility. Economic Journal, 101, 483-496.
Blanchflower, D. G. and Oswald, A. J. (2004). Happiness and the human development index: the paradox of Australia. The Australian Economic Review, 38(3), 307-318. Blanchflower, D. G. (2007). Is unemployment more costly than inflation?. National
Bureau of Economic Research, 1-40.
Clark, A. E., & Oswald, A. J. (1994). Unhappiness and unemployment. The Economic
Journal, 648-659.
Clark, A. (2003). Inequality-aversion and income mobility: a direct test. Delta.
Di Tella, R., MacCulloch, R. J., & Oswald, A. J. (2001). Preferences over inflation and unemployment: Evidence from surveys of happiness. American Economic Review, 335-341.
Di Tella, R., MacCulloch, R. J., & Oswald, A. J. (2003). The macroeconomics of happiness. Review of Economics and Statistics, 85(4), 809-827.
Di Tella, R.and MacCulloh, R. (2006). Social uses of happiness data in economics. The
Journal of Economic Perspective, 20(1), 25-46.
Diener, E., Diener, M., & Diener, C. (1995). Factors predicting the subjective well-being of nations. Journal of Personality and Social Psychology, 69, 851-864.
Diener, E., Suh, E. M., Lucas, R. E., & Smith, H. L. (1999). Subjective well-being: Three decades of progress. Psychological Bulletin, 125(2), 276.
29
Easterlin, R. A. (1974). Does economic growth improve the human lot? Some empirical evidence. Nations and Households in Economic Growth, 89, 89-125.
Easterlin, R. A. (1995). Will raising the incomes of all increase the happiness of all?.
Journal of Economic Behavior and Organization, 27(1), 35-47.
Ferrer-i-Carbonell, A. (2013). Happiness economics. SERIEs, 4(1), 35-60.
Frank, R. H. (1985). Choosing the right pond. New York and Oxford: Oxford University Press. Frey, B. S. and Stutzer, A. (2002). What can economist learn from happiness research?.
Journal of Economic Literature, 40, 402-435.
Graham, C., & Sukhtankar, S. (2004). Does economic crisis reduce support for markets and democracy in Latin America? Some evidence from surveys of public opinion and well being. Journal of Latin American Studies, 36(02), 349-377.
Graham, C. (2005). The economics of happiness: Insights on globalization from a novel approach. World Economics, 6(3), 41-55.
Graham, C. and Felton, A. (2006). Inequality and happiness: insights from Latin America. Journal of Economic Inequality, 4, 107-122.
Grosfeld, I. and Senik, C. (2010). The emerging aversion to inequality. Economics of
Transition, 18(1), 1-26.
Hall, R. E. and Taylor, J. (1997). Macroeconomics, 5th Ed. New York: Norton Press.
Helliwell, J. F. (2003). How’s life? Combining individual and national variables to explain subjective well-being. Economic Modelling, 20, 331-360.
Hirsch, F. (1976). The social limits of growth .Cambridge, Mass: Harvard University Press. Inglehart, R. (1990). Culture shift in advanced industrial society. Princeton University
Press.
Inglehart, R., Foa, R., Peterson, C., and Welzel, C. (2008). Development, freedom, and rising happiness: A global perspective. Perspective on Psychological Science, 3(4), 264-285.
Mankiw, N. G. (1997). Macroeconomic, 3rd Ed. New York: Worth Publishers.
Marks, G. N. and Fleming, N. (1999). Influences and consequences of well-being among Australian young people: 1980–1995. Social Indicators Research, 46(3), 301–23.
30
McBride, M. (2010). Money, happiness, and aspirations: Am experimental study. Journal
of Economic Behavior and Organization, 74: 262-276.
Michalos, A. C. (1991). Global report on student well-being (Vol. 1). New York: Springer. Osberg, L. (2014). Measuring economic insecurity in rich and poor nations. Review of
Income and Wealth, 1-24.
Oswald, A. J. (1997). Happiness and economic performance. The Economic Journal, 107, 1815-1831.
Peiro, A. (2006). Happiness, satisfaction and socio-economic conditions: Some international evidence. The Journal of Socio-Economics, 35, 348-365.
Scitovsky, T. (1976). The joyless economies. Oxford: Oxford University Press.
Shiller, R. J. (1997), ‘Why do people dislike inflation?’ in Christina Romerand David Romer (eds.), reducing inflation: motivation and strategy. Chicago: University of
Chicago Press, 13–65.
Stock, J. H., & Watson, M. W. (2012). Introduction to econometrics. Boston: Addison
Wesley.
Veenhoven, R (2014). World database of happiness. Erasmus University Rotterdam, The
Netherlands.
Winkelmann, L. and Winkelmann, R. (1998). “Why are the unemployed so unhappy? evidence from panel data,” Economica. 65 (257), 1–15.
Wolfers, J. (2003). Is business cycle volatility costly? Evidence from surveys of subjective well-being. International Finance, 6(1), 1-26.
