There is no doubt that the positive or negative view that we have of our lives is of huge importance to each of us. For this reason economists frequently recommend to include measures of subjective well-being in measures of social progress – and in particular to augment the usual measure of economic prosperity (GDP per capita).1
‘Overall happiness is the degree to which an individual judges the overall quality of his/her own life-as-a-whole favorably. In other words: how much one likes the life one leads.’ This is the definition given by the widely used ‘World Database of Happiness’ (here).
# Empirical View
# Comparing Happiness across Countries
The map below presents data from the World Happiness Report 2016 – online here. Click on ‘Chart’ to see the change over time for a particular country.
# Correlates, Determinants, and Consequences
# Life satisfaction and income
# Life satisfaction and income across Countries
One of the most influential recent study was done by the Princeton professor Angus Deaton in 2008. He had the enviable opportunity to study a dataset of life satisfaction that was constructed by the Gallup Organization. This data comes from an identical questionnaire that was used to ask people in 132 countries of the world about theirlife satisfaction.2
In the graph reprinted below, Deaton plotted the way that people view their lives versus their income. The result is very clear and shows a strong positive correlation between economic prosperity and people’s view of their own lives.
Global data on the correlation between income and life satisfaction from Gallup World Poll – Deaton (2008)3
# Interactive chart: Life satisfaction and income across Countries
The scatterplot shows the same link between life satisfaction and income across countries for a different dataset on life satisfaction.
# Life satisfaction and income within countries
The correlation between incomes and life satisfaction is not just high across countries but also within countries. The same Gallup data that Deaton used was analysed by Betsey Stevenson and Justin Wolfers who studied the relationship between life satisfaction and income within each country. The correlation coefficients (this statistical measure is explained in Wikipedia) between life satisfaction and income are shown in the following graph – they are positive in all but one country and indicate that higher incomes go together with higher life satisfaction.
Distribution of estimates of the within-country life satisfaction-income gradient from Gallup World Poll – Stevenson & Wolfers (2008)4
# Life satisfaction and income within and across countries
So far we have seen that life satisfaction differs widely between different countries and that this variation can be explained in the differences of economic prosperity of the country. This link between income and life satisfaction is true across and within countries.
All of this information is successfully shown in a single graph by Stevenson and Wolfers. It is again visualizes the Gallup World Poll data – the black solid circles show the correlation between life satisfaction and income across countries, the slope of the arrow represents the correlation coefficients that measures how strongly incomes and life satisfaction are associated within each country.
Within-country and between-country estimates of the life satisfaction-income gradient from Gallup World Poll – Stevenson & Wolfers (2008)5
Source of the data on subjective well-being is the Gallup World Poll, 2006.
Note to the Graph from the paper: ‘Each solid circle plots life satisfaction against GDP per capita for one of 131 developed and developing countries. The slope of the arrow represents the satisfaction-income gradient estimated for that country from a country-specific ordered probit of satisfaction on the log of annual real household income, controlling for gender, a quartic in age, and their interaction. Usable household income data were unavailable for eighteen countries. The dashed line represents the between-country satisfaction-income gradient estimated from an OLS regression of the satisfaction index on the logarithm of real GDP per capita. GDP per capita is at purchasing power parity in constant 2000 international dollars.’[/ref]
# Early studies on the link between life satisfaction and income
The result that Deaton obtained is replicated in other studies using other data. In their influential paper, Stevenson and Wolfers compared different datasets and confirmed the robustness of the result described above.
The following graph shows the correlation of measures of subjective well-being and income for five very early data sets.
Early cross-country surveys of subjective well-being for different samples, 1946-1965 – Stevenson & Wolfers (2008)6
The sources of data for these graphs are Cantril (1951); Buchanan and Cantril (1953); Strunk (1950); Cantril (1965); Veenhoven (undated); Easterlin (1974, table 7); Maddison (2007). For more information on the shown correlations see the original paper.[/ref]
# Living conditions and health
# Life satisfaction and child mortality
The following visualization shows the lower self-reported happiness level in countries in which the population suffers from poor health, here measured as high child mortality.
# History and Culture
Another recent confirmation of the huge variability of subjective well-being across countries and its strong link with income is presented by researchers working on the World Value Survey. Their result is plotted below. Note that the income scale on the x-axis is not logarithmic as in the graph shown before, for this reason the correlation of subjective well-being appears to be flat for high income countries.
The black line shows the average relation between income and subjective well-being and income. The fact that similar countries lie collectively above or below this line shows that income – and its many correlates – are not the only determinant of subjective well-being. The cultural background might explain why people in the culturally similar countries of Latin America have a higher subjective well-being than their comparably low income would predict.
The shared experience of the grim communist era in many countries might explain why the subjective well being of the people in the ex-communist countries is lower than their income would predict. But again for both samples the correlation between incomes and well-being is intact.
Subjective well-being (SWB), per capita gross domestic product (GDP) and different types of societies – Inglehart, Foa, Peterson, and Welzel (2008)7
The paper can be freely downloaded from the ‘World Value Survey Website’ here.[/ref]
# Comparing Life Satisfaction through Time
The relationship between income and happiness/life satisfaction over time appears to be less definite when we do not look at the correlation across countries but within the same country over time. The famous result of the early literature on this relationship was that there is no link; this result is called the Easterlin Paradox, named after the economist Richard Easterlin who published this result of a non-link between the two variables.8 In 2010 Easterlin restated his paradox.9
It is important to be clear that Easterlin is not saying that the results shown above are wrong. He is not saying that there is no relation between income and happiness across countries. It is precisely the observation that there is a strong link across countries why he speaks of a paradox – in the 2011 article he says: ‘If there were no positive relation in the cross-section, there would be no paradox!’
The time series of measured subjective well-being in Japan over three decades of strong economic growth in Japan is shown below – and is taken from a publication of Easterlin.
Mean subjective well-being in Japan, 1958-1987 – Easterlin in Land, Michalos, and Sirgy (ed.) (2011)10
In the remainder of this entry I want to look at the inter-temporal data on life satisfaction. Because incomes have been growing over the last century in most countries, if we claim incomes and life satisfaction do not go together, it suffices to say that it is life satisfaction that is not rising.
Comparisons of life satisfaction over time are only possible for developed countries. In developing countries researchers have only recently started to measure life satisfaction, so an inter-temporal perspective cannot be assumed.
Of all the developed countries, the most often cited case as a proof of the Easterlin Paradox is the case of Japan. The data for these claims are from the ‘Life in Nation’ surveys that started to measure life satisfaction as early as 1958 in Japan. Japan is a very interesting case because it had very rapid economic growth over the post-war period until the early 1990s. The growth miracle of Japan meant average growth rates of more than 5% over the 1960s, 70s and 80s.
How did this affect life satisfaction of the Japanese? Previous researchers looked at the data in the ‘Life in Nation’ surveys uncritically and concluded that life satisfaction did only oscillate around the long term average, and over the long run the Japanese became neither substantially more nor substantially less satisfied.
Stevenson and Wolfers whose research I quoted before examined the Japanese data over the last 50 years very carefully and translated all of the survey questions. One example of how the questions changed: For the first 6 years the survey was kept consistent, but then in 1964 the top category of the responses was changed from the catch-all, “Although I am not innumerably satisfied, I am generally satisfied with life now” to the much more narrow “Completely satisfied.” Therefore it does not come as a surprise that the proportion reporting their well-being in this highest category dropped from 18.3% to 4.4%.
Stevenson and Wolfers analyze all surveys from 1958 to 2007 and find that ‘properly viewed, [this analysis] leaves us with four periods within which we can make useful assessments of trends in subjective well-being in Japan’.
These results, together with short summaries of how the survey questions changed, are shown in the following graph.
Changing interview questions over time and their effect on the relation between measured life satisfaction and GDP per capita in Japan – Stevenson & Wolfers11
Data source of the graph’s authors: Life in Nation surveys, 1958–2007.
Note from the paper: ‘The series in each of the four panels reports responses to a different life satisfaction question, and therefore comparisons should be made only within each panel. GDP per capita is at purchasing power parity in constant 2000 international dollars.’[/ref]
Treating the surveys of Japan as the four distinct datasets that they really are, the authors find what we would expect on the basis of the cross-sectional data: life satisfaction rises with rising incomes (until the early 1990s), and over the so called ‘lost decade’ of the 1990s and the stagnating incomes of the recent past life satisfaction is falling.
The authors’ argument is compelling in light of their examination of the surveys: ‘By construction, the levels of these four series are not comparable, and hence comparisons within, but not between, series are valid.’ From their careful analysis summarized in the graph above, we learn that life satisfaction rises as economic growth lifts societies out of poverty. The conclusion is that Easterlin’s Paradox should be more seen as an important warning to be critical with data sets that combine data points without checking if this data can be combined. This example makes clear why the research with identical questionnaires in different countries is so insightful for our understanding of how poverty and low life satisfaction and prosperity and high life satisfaction are related.
It might also be relevant that compared with the other countries the Japanese are experiencing a lower life satisfaction than you would predict given their high income – note that Japan scores lower than the black line in the graph in the previous section.
Having this warning in mind, I still want to cautiously refer to one study of inter-temporal comparisons of happiness. The ‘World Value Survey’ carries out polls across the globe since the 1980s. Based on this research Inglehart, Welzel and Foa looked at the changes of both happiness and self-assessed freedom of choice. The strong correlation between both measures and the trend towards a freer and happier world is shown in this last graph of this entry.
Changes in subjective well-being and sense of free choice – Inglehart, Foa, Peterson, and Welzel (2008)12
The paper can be freely downloaded from the ‘World Value Survey Website’ here. [/ref]
# Data Quality & Definition
The main reason why direct measures of personal well-being are not more commonly used is that it is very hard to measure.13
# Measurement and Data
A major distinction should be made between positive or pleasant emotions (happiness) and well-being and life-satisfaction. The first is concerned with the experience of positive emotions at any given time, the second is concerned with rather long-term evaluation of one’s own life – it is thus closer to view that philosophers and religious thinkers often take. Normative concepts of the good life matter more for this self-assesment than momentary emotions.
Measurement of the two components of the experience of one’s own life is typically done by interviews and surveys.
# Problems of Happiness Assessments
There are two main problems:
Across countries: Cultural and linguistic differences might mean that the same survey questions are viewed differently by people from different cultural backgrounds and depending on the language.
Across time: Comparisons of happiness measures over time unfortunately should be checked especially critically. The reason is that it is very often the case that the questions are changed and that the changes of happiness over time reflects less the change of happiness of the people and rather changes of the assesment by which the happiness is measured.
# Data Sources
The World Database of Happiness (hosted at the ‘Erasmus University Rotterdam’ here) encompasses a wealth of data and study results on happiness – from the national to the global level across. It includes measure of inequality of happiness. And there are many time series for a huge range of countries. You also find correlational studies with just about measure you can imagine.
A range of important measures (including links to the sources) is described in Wikipedia here.