Why am I interested in happiness research?

“THOMAS GRADGRIND, sir. A man of realities. A man of facts and calculations. A man who proceeds upon the principle that two and two are four, and nothing over, and who is not to be talked into allowing for anything over. Thomas Gradgrind, sir - peremptorily Thomas - Thomas Gradgrind. With a rule and a pair of scales, and the multiplication table always in his pocket, sir, ready to weigh and measure any parcel of human nature, and tell you exactly what it comes to. It is a mere question of figures, a case of simple arithmetic. You might hope to get some other nonsensical belief into the head of George Gradgrind, or Augustus Gradgrind, or John Gradgrind, or Joseph Gradgrind (all supposititious, non-existent persons), but into the head of Thomas Gradgrind - no, sir!

In such terms Mr. Gradgrind always mentally introduced himself, whether to his private circle of acquaintance, or to the public in general.” Charles Dickens, Hard Times, Chapter2.

I'm not like that. However, an interest in happiness research may seem to many people to imply an obsession with measuring, calculating and attempting to understand things that are not meant to be understood. 

Perhaps trying to understand what makes people happy is a bit like trying to understand why jokes are funny. It isn’t obvious that an understanding of what makes jokes funny would be much help to anyone in telling jokes, or how an understanding of what makes people happy would help anyone to become happier.

It is fairly easy to explain how I came to be interested in happiness research, so I will begin by writing about that. In my work as an economist I spent more than a few decades considering what government policies were likely to advance the well-being of the people in the countries where I have lived and worked (mainly Australia and New Zealand. It seemed fairly obvious that the vast majority of Australians and New Zealanders wanted higher incomes, so it was reasonable to assume that would improve their well-being. If someone questioned whether higher incomes would make people any happier, my defence was that economists should be in the business of making it possible for people to have happier lives rather than advising them how to spend their money.

At the same time, I could not help becoming interested in the puzzle of why happiness surveys showed little or no increase in average happiness ratings in high income countries over several decades while average income levels rose substantially. This is of course Easterlin’s puzzle - named after the economist Richard Easterlin.

I stopped being puzzled once I understood that happiness surveys measure emotional well-being - a component of well-being rather than the whole package. There is no reason to expect the value that people place on physical health, education, housing and safety, among other things, to be fully reflected in measures of emotional well-being. 

Emotional well-being is strongly related to self-esteem, optimism and the feeling of being in control of one’s life – none of which would be expected to be strongly influenced by further increases in incomes in high-income countries.

It is true, of course, that when people see higher incomes as the pathway to emotional bliss they are unlikely to be satisfied with one pot of gold - even if they find the end of a rainbow. But most people seem to make sensible choices. They might seek a higher income if that is necessary to pursue objectives that they consider to be worthwhile. For many people, higher incomes are incidental to career objectives. There is no reason to expect people to stop trying to achieve more in life just because they are satisfied with their current standard of living.

It seems to me that if we are interested in measuring well-being, then the survey measures of happiness are just one of the items we should look at. I favour the approach taken by the OECD in its Better Life Index.

However, an indicator approach doesn’t give economists a value-free measure of well-being. It leaves open the question of what weights should be given to the various component indexes. The OECD leaves the value judgement in the hands of the users of its index. That is more appropriate than having researchers assign weights, but it would be good to see how weights might need to differ to reflects the different values of people in different parts of the world. In my view the Better Life Index should be accompanied by illustrative weights derived from a values survey.

So, one of the reasons why I am interested in happiness research is apparent from what I have written. Happiness research is relevant to measurement of human well-being and that is relevant to economic policy.

I am particularly interested in the relationship between freedom and flourishing. Do government restrictions on individual freedom – in the wars against drug taking, smoking, alcohol, obesity, overwork etc. - actually have the desired effect of enabling people to have happier lives? I don’t think so. The policies adopted by governments seem designed to make people less happy in an attempt to get them to adopt healthier lifestyles, but I don’t know where to find the evidence to prove it.

In any case, that is only part of the story. My interest in happiness research is not always closely related to government policy. Some of my recent posts have taken me into the relationship between life satisfaction and the incidence of negative emotional experience. I am not sure why I am interested in such matters. Nevertheless, it seems more satisfying than spending my time trying to understand what makes jokes funny.

Is the incidence of depression higher where people are experiencing more negative emotion and suffering?

It would be difficult to be interested in human flourishing without having some interest in understanding the differences in incidence of “suffering”, negative emotion, and depression in different parts of the world.

I have written a few posts about negative emotion recently, but it is now a few years since I looked explicitly for a relationship between “suffering” and depression at an international level. The last time I looked, I couldn’t see any relationship. I concluded that “the data … suggest there is no simple relationship”.

There is now some new data available, so I have taken another look. But I don’t want you to get too excited, because you might end up feeling disappointed.

A data set on depressive disorders in 2010 has been compiled by Alize Ferrari and others as part of the Global Burden of Disease Study. The researchers collected data on the depressive disorders from published articles and pooled the data using a statistical technique (Bayesian meta-regression). The data used below relates to major depressive disorder (MDD) which is also known as clinical depression. It involves at least one major depressive episode in which the affected individual experiences a depressed mood almost all day, every day for at least 2 weeks.

The negative emotion data is sourced from the Gallup World Poll. Respondents were asked if they experienced worry, anger, sadness or depression yesterday. The data used is based on country averages of yes/no answers (yes = 1).

The data on suffering is from the new Gallup-Healthways Global Well-being index. The index covers 5 dimensions of well-being: purpose (liking what you do and learning or doing something interesting); social (having supportive relationships); financial (having enough money to do what you want and not being worried about money); community (liking where you live and having pride in your community) and physical (feeling active and productive, and in good health). People are classed as suffering in a particular element if their well-being in that element is low and inconsistent.

For the purpose of this exercise I have estimated the percentage suffering in each country as the average of the percentages suffering in each dimension.

After the three data sets were combined I had data for 114 countries. At the country level, there is a modest degree of correlation between the three variables:

Struggling and MDD:                          0.24

Struggling and negative emotion:        0.20

Negative emotion and MDD:              0.06 .

The graph presented below provides a basis to compare averages for the relevant variables in 10 regions of the world.

All three indicators tell an unambiguous story of misery in the Middle East and North Africa. Suffering seems somewhat higher in Sub-Saharan Africa, but the incidence of negative emotion is apparently lower in that part of the world than in the Middle East and North Africa.

Central and Eastern Europe seems to have a fairly high incidence of all three indicators of misery. The incidence of depression and suffering is similarly high in the former Soviet countries, but the incidence of negative emotion is similarly lower.

East Asia stands out as having a relatively low incidence of negative emotion and depression, but a higher incidence of suffering than Latin America. According to the Gallup-Healthways data, the incidence of suffering in Latin America is not much greater than in Europe.

It is possible to speculate on reasons why the three indicators tell different stories in several regions of the world. Cultural factors probably explain the relatively low incidence of negative emotion in East Asia and Sub-Saharan Africa. Cultural factors might also explain why the incidence of suffering has been assessed to be relatively low in Latin America. (As discussed in an earlier post, cultural factors might also explain why the Gallup-Healthways index shows a high percentage of the population in Latin America to be thriving). There may also be potential for the estimates of depression to be biased by such factors as differences in availability of trained mental health professionals.

That leaves me concluding that caution is required in comparing suffering, negative emotion and depression in different parts of the world. I know that is a fairly lame conclusion, but I did warn you not to get too excited.

Do international comparisons show that people in countries with low average life satisfaction tend to have a high incidence of negative emotion?

It seems intuitively reasonable that people in countries with low average life satisfaction levels would tend to have a higher incidence of negative emotion. The theory of subjective well-being homeostasis, discussed in my last post, provides a theoretical basis to predict that will happen. Adaptation and resilience normally keep life satisfaction within a set-point range, but if resilience is weak, life satisfaction can fail to recover from negative experiences. On that basis we would expect low average life satisfaction to be associated with a relatively high incidence of homeostatic breakdown.

At first sight, Figure 1 appears to provide very limited support for the homeostasis theory. There are no countries in which high average life satisfaction is accompanied by a high incidence of negative emotion. At the other end of the scale, however, there are many countries in which low average life satisfaction is accompanied by a relatively low incidence of negative emotion.

The data is sourced from the Gallup World Poll (via World Happiness Report 2013). The negative emotion data for each country is the average of yes/no answers (yes = 1) to the question of whether respondents experienced worry, anger, sadness, anger and depression yesterday. The life satisfaction data is based on the Cantril ladder which involves survey respondents being asked to rate their lives against an 11 point scale in which the top rung of the ladder (rating of 10) corresponds to the best possible life and the bottom rung of the ladder (rating of 0) corresponds to the worst possible life.

Figure 2 shows the expected relationship after controlling for a range of socio-economic and cultural factors. This involved adjusting the data on incidence of negative emotion using the results of a regression analysis. The adjusted data are estimates of what the incidence of negative emotion might have been in the absence of variation in the socio-economic and cultural factors (i.e. with the socio-economic variables equal to the average over all countries and European/American culture).

The regression analysis suggests that at a national level an increase of 1 unit in average life satisfaction reduces the incidence of negative emotion by about 0.02 (SE = 0.007) i.e. by about 10% at the world average level of negative affect. The regression explained about 45% of international variation in the incidence of negative emotion.

The socio-economic variables included in the regression were per capita GDP, social support (relatives and friends to count on), freedom (proportion satisfied with freedom to choose what they do) and corruption (proportion saying corruption is widespread in business or government). The estimated coefficients for those variables were all significantly different from zero, with a negative estimated coefficient on income. It isn’t surprising that high average incomes could be associated with a high incidence of negative emotion if not accompanied by high average life satisfaction and social support.

The cultural influence has been accounted for by using regional dummy variables. The estimates suggest that cultural factors reduce the reported incidence of negative emotion by the following amounts:

East Asia:                                 0.134 (SE 0.025)

Africa:                                      0.104 (SE 0.017)

South Asia:                              0.103 (SE 0.027)

Former Soviet Union:              0.092 (SE 0.020)

Central and Eastern Europe:    0.055 (SE 0.018)

South East Asia:                       0.050 (SE 0.022)

The low incidence of reported negative emotion in East Asia is consistent with previous research on cross-cultural difference in subjective wellbeing. (See, for example a recent article by Lufanna Lai, Robert Cummins and Anna Lau – abstract here.)

One of the most interesting findings of the regression analysis reported above is that the coefficient for Latin America was not significantly different from zero. This is in contrast to the findings of studies relating to positive emotion (including those reported on this blog here and here) which suggests that Latin American culture has a strong positive impact. It seems that the positivity of Latin Americans does not translate to a lower incidence of negative emotion in that part of the world. 

Does homeostasis explain the stability of life satisfaction in high income countries?

The theory of subjective well-being homeostasis proposes that life satisfaction is controlled by automatic neurological processes in a manner analogous to the maintenance of body temperature. The theory has been proposed by Professor Robert Cummins of the Australian Centre on Quality of Life (ACQOL), Deakin University. The basic idea is that positive and negative experiences can cause temporary changes in life satisfaction, but homeostasis normally brings it back within a set-point range. The exception occurs where chronic failure of the homeostatic system results in depression.

The theory proposes that homeostatic buffers enable humans to function normally:

• adaptation restores life satisfaction to the set-point range following positive changes in circumstances, such as an increase in income;
• resilience tends to restore life satisfaction to the set-point range following strong negative challenges.

Resilience depends on external resources, particularly intimate relationships and wealth, and internal buffers designed to minimise the impact of personal failure on positive feelings about the self. The internal buffers can operate at an unconscious level - e.g. assisting an individual to cope by enabling positive emotions to become accessible. They also operate at a conscious level by altering the way individuals see themselves in relation to the challenge (e.g. denying personal responsibility or viewing the failure as unimportant).

The distinguishing feature of the homeostasis theory is not the existence of adaptation and resilience – which are widely acknowledged in the happiness research literature – but the idea that these processes tend to restore emotional systems to unique set-points for each individual.

When I first read about homeostasis theory, a few years ago, evidence that some individuals experience long-term changes in life satisfaction seemed inconsistent with the idea of constant individual set-points. However, as Bob Cummins has pointed out, changes in life satisfaction do not necessarily imply change in an individual’s set-point (or homeostatically protected mood). If initial measurements of life satisfaction are higher than, or lower than, the set-point, then subsequent measurements can be expected to show a return to the set-point range.

My remaining doubts about homeostasis theory centre mainly around the question of how it can be reconciled with the international evidence of lower average life satisfaction in low-income countries. I find it hard to accept that a high proportion of the people in low-income countries who claim to have relatively low levels of satisfaction with their lives are suffering failure of their homeostatic systems. In some low-income countries, e.g. China, relatively low average life satisfaction seems to be accompanied by relatively high positive affect and relatively low negative affect.

Perhaps set-point ranges remain constant – if we think in terms of hypothetical neural correlates – but the relationship between set-point ranges and life satisfaction scales may change with changes in perceptions of what might be possible. As discussed in a recent post, individuals who move from a remote village to a major city might feel that their lives have improved, even though they become less contented with their living standards after moving to the new location. Does that mean they have become more vulnerable to homeostatic breakdown? I am not qualified to make informed predictions in relation to such matters, but my guess is that there would not be an increased risk of breakdown if the people concerned remain optimistic about their prospects in the city and retain the option to return to the village for family support if they need it.

The predictions of homeostasis theory seem to stand up well to tests that have been conducted so far. For example, homeostasis theory predicts that there will be greater variation in subjective well-being among people with low incomes than among people with high incomes. This is because people with low incomes (or low wealth) are more vulnerable to changes in circumstances. Analysis using data from the Australian Unity Wellbeing Index indicates that the standard deviation of subjective well-being is indeed substantially higher for people with relatively incomes, and declines as household incomes rise to around $100,000.

Homeostasis theory seems to provide a more plausible explanation for the stability of average life satisfaction in high-income countries than a rival theory, advanced by some economists, that this stability reflects evaluative judgements of life by the people in those countries. Evidence that life satisfaction is influenced by genes, and strongly related to self-esteem, optimism and feelings of being in control of one’s life, suggests that it is more appropriately interpreted as reflecting the moods (or frames of mind) of respondents than evaluative judgements. 

Researchers who want subjective well-being measures to reflect evaluative judgements that individuals make about the quality of their lives should consider an approach which requires greater cognitive inputs e.g. the ACSA question which asks people to assess their current well-being relative to the best and worst periods of their lives.

Is Australia's political system broken?

“The process of debate, competition and elections leading to national progress has broken down. The business of politics is too de-coupled from the interests of Australia and its citizens. This de-coupling constitutes the Australian crisis”.

The quote is from the last chapter of Paul Kelly’s most recent book, Triumph and Demise, The broken promise of a Labor Generation (published by MUP, but I purchased the Kindle version from Amazon). Over several decades Paul Kelly has developed a formidable reputation as Australia’s leading ‘big picture’ journalist. I don’t think anyone else has been more successful in identifying the defining elements of the changing political landscape in Australia.

In launching the book, Tony Abbott, Australia’s current prime minister, took issue with Kelly’s claim that the system has broken down. He suggested that the style of government of the past six years had been just a passing phase, rather than “the new normal”. The PM said:

“It’s not the system which is the problem, it is the people who from time-to-time inhabit it. Our challenge at every level is to be our best selves.”

More recently, the electoral triumph of the New Zealand National Government under John Key appears to have demonstrated that it is still possible for a government to contain the growth of government spending in the face of political obstacles similar to those faced by governments in Australia.

So, is Paul Kelly wrong? I don’t think so. The system must be broken if achievement of reforms requires politicians to become their “best selves”.

Paul Kelly writes of “volatility and fragmentation” as being “the new driving forces” of Australian politics. He argues that the system has evolved in ways that have given sectional interests more power than ever before. He mentions technology and campaign techniques in this context, and brings fragmentation of the traditional media and the rise of social media into the discussion. He also makes the point that it has become more difficult for leaders to talk honestly to the community as they have become subjected to greater media pressure to rule out any action that might disadvantage any powerful interest group.

However, the strongest points which Paul Kelly makes in support of the view that sectional interests now have more power relate to the rise of minor parties, accompanied by a decline in tribal loyalty to the major parties. Minor parties tend to play to sectional interests because “they lack any governing culture or responsibility for the national interest”.

It has for many years been a normal part of the Australian political landscape for government legislation to be subject to obstruction from the cross-benches in the Senate. In the past, Governments were often able to deal effectively with this by threatening a double dissolution election. Paul Kelly points out that following the expansion in size of Parliament in the mid-1980s, there is now more likelihood that a double dissolution election will result in even more minor party representation because the quota of votes required for election is easier to obtain.

It is possible, of course, for the major parties to come together to agree to the spending cuts or tax increases required to avoid an explosion in government debt. But it is particularly difficult for the major parties to be seen to agree on a strategy to achieve that. The small remaining philosophical differences between the major parties hinge around fiscal policies - the Labor party tends to favour a bigger role for government, while the Liberal party tends to be somewhat opposed to government playing a bigger role.  In addition, even though there may not be much difference in practice between the policies the major parties implement when they are in government, when in opposition neither party has an incentive to offend interest groups that might help them win to the next election.

Perhaps the current government will be able to resolve the immediate fiscal problem by proposing spending cuts that will be more acceptable to the opposition, or the minor parties, than those currently proposed. The chances of such reforms being accepted will improve as the next election approaches and the opposition comes under greater pressure to demonstrate fiscal responsibility. As government debt increases, it will also become increasingly difficult for the opposition to oppose fiscal reform on the grounds that Australia is not Greece.

However, even if the immediate fiscal problem can be resolved satisfactorily, it is difficult for anyone to claim that Australia’s political system is not broken when it contains in-built incentives for fragmentation of the major parties. As the minor parties demonstrate to interest groups the advantages of Senate representation, voters have a strong incentive to break their allegiance to the major parties. This is particularly likely if (to use words of my friend Jim Belshaw on his blog a few years ago) they “reject the intellectual, institutional and political constructs” of the major parties. As the parties fragment, some excellent minor party candidates, like Senator David Leyonhjelm of the LDP (whom I claim to have voted for intentionally in the last election) will probably be elected.  In the end, however, we are likely to be left with a situation where the line of least resistance will usually prevail and governments will always be able to blame Senate obstruction for poor economic policy outcomes.

In an article posted here a couple of months ago, I suggested that better policy outcomes could be expected if Senate candidates were chosen randomly, rather than selected by parties. There is not much chance of a proposal along those lines being seriously considered in the near future.

There are probably other ways to reduce Senate obstruction that would be more acceptable to the major parties. The major parties will be doing a useful service to themselves, as well as the nation, if they can get together to consider and implement reforms to the Senate before the benefits of accountable government disappear entirely. 

Postscript: 
Jim Belshaw has responded to this post on his blog.

What does it mean to be thriving?

I am asking myself that question because I am trying to come to grips with the findings of the new Gallup-Healthways Global Well-Being Index.

The methodology of the index looks like a sensible way to assess the extent to which people are thriving in different countries. Surveys are used to obtain subjective data relating to the following five elements:

• Purpose: liking what you do every day and learning or doing something interesting every day
• Social: having supportive relationships - someone who encourages you to be healthy and family and friends who give you positive energy
• Financial: having enough money to do what you want to do and not being worried about money
• Community: liking where you live and having pride in your community
• Physical: feeling active and productive, and that your physical health is near perfect.

When I look at the results, however, I wonder whether the new Gallup-Healthways index actually measures the extent to which people are thriving.

The top ten countries in the index are as follows:

1. Panama
2. Costa Rica
3. Denmark
4. Austria
5. Brazil
6. Uruguay
7. El Salvador
8. Sweden
9. Guatemala
10. Canada.

When Scott Sumner looked at that ranking he wasn’t surprised to see some of the countries of Latin America do well, but he was shocked to see Sweden bracketed by Guatemala and El Salvador.  His comment:

“Just to be clear, I’m NOT saying that the people in those two countries are not just as happy as the Swedes; for all I know they are happier.  I have no idea how to measure happiness. But if you are talking about country rankings, people are going to assume you are making some sort of statement about socio-economic/political systems.  And if a large share of the people in these highly successful societies are risking murder, rape and dying of thirst in order to flee to a country where they don’t speak the language, so that they can get jobs cleaning toilets or picking vegetables in the hot sun all day long, then I have to wonder whether these rankings actually mean much of anything”.

Gallup’s potential net migration index suggests that large numbers of people who live in Guatemala and El Salvador would indeed prefer to live elsewhere. Recent surveys suggest that while about 28% of the population of Guatemala would prefer to live elsewhere, the corresponding figure for El Salvador is about 33%.

In Free to Flourish I made the point that if you want to measure the quality of different societies it makes more sense to attempt to define the characteristics of a good society and attempt to measure the extent to which societies have those characteristics, rather than to attempt to infer the quality of a society solely from happiness indexes. Nevertheless, it comes as a surprise when a high proportion of the population is assessed to be thriving in societies from which large numbers of people wish to migrate to seek better opportunities.

When I set out to find out the reasons for the results obtained by Gallup-Healthways my first thought was that it might reflect the method used to rank countries. The criterion used is the percentage of the population that are thriving on the basis of three or more of the elements defined above. However, when I constructed an index by averaging the scores on all five elements (giving thriving a rating of 3, struggling a rating of 2 and suffering a rating of 1), El Salvador remained in 7^th place and the ranking of Guatemala remained fairly high (falling from 9^th to 14^th).

My second thought was that people would be unlikely to give equal weight to the five specified elements in assessing the quality of their lives. In order to assess what weights might be appropriate I used regression analysis to explain the old Gallup thriving index in terms of the five elements of the new index. The old Gallup index is based on the Cantril methodology under which survey respondents are asked to evaluate their own lives relative to the best and worst possible life. Under the old Gallup index the percentages of the population assessed to be thriving in El Salvador and Sweden were 36% and 68% respectively.

There is a problem with the use of regression analysis to obtain weights because the old and new indexes relate to surveys taken years apart, but that seemed to me to be a minor problem by comparison with use of equal weights.

The results of the analysis suggest that it might be appropriate to give a weighting of 40% to Purpose, 30% to Financial, 30% to Physical, and zero weight to Social and Community. The rankings on that basis are:

1. Panama 
2. Sweden
3. Denmark
4. Austria
5. Costa Rica
6. Canada
7. Netherlands
8. Iceland
9. Mexico
10. El Salvador

At this point I have to acknowledge that the high rating given to El Salvador is unlikely to be a result of the methodology used for ranking, or failure to weight elements appropriately.

No matter how I look at it, the people of El Salvador seem to be highly positive about their lives. This is consistent with the results of other Gallup surveys which have shown that the people of El Salvador experience a great deal of positive emotion.

The problem is that while it is good to have positive emotional states or positive states of mind, thriving involves more than that. From observed behaviour it is obvious that humans see their ability to thrive as related to objective circumstances such as incomes, life expectancy and education – which are reflected in the UN’s Human Development Index (HDI) – as well as to their emotional states.

Anyone interested in identifying the countries in which people have the best opportunities to thrive might find the following chart of some interest.