Because more equal societies work better for everyone

Frequently Asked Questions

To suggest that these relationships are causal does not involve a major departure from what we know already.  Within countries we know that all the components of our Index of Health and Social Problems are strongly related to social status: the further down the social ladder the more common they become.  The new part of the picture is simply that if you stretch out the social status differences all the problems related to social status become more common. Rather than postulating entirely new causal processes we are therefore only providing a bit more information about the relationships that have always been recognised.

People who have studied the graphs on this web site and in The Spirit Level carefully will have noticed that there is a clear tendency for countries which do badly on one outcome to do badly on others.  We show evidence that 10 or 12 different problems tend to move together. That implies that they share an underlying cause.  The association between inequality and our Index of Health and Social Problems is very close and no one has yet suggested an alternative. 

Lastly, as the different chapters in our book show, many of the causal processes leading from inequality to the various health and social problems are already known. For example, the effects of social status on health have been demonstrated among monkeys in experiments which kept diet and material conditions the same while altering social status by moving animals into new groups and the effects of chronic stress on the immune and cardiovascular systems are increasingly well understood.  Similarly, violence is more common in more unequal societies (where status competition is intensified) because it is so often triggered by people feeling looked down on, disrespected and humiliated

As the book's intended readership was not confined to those with statistical training, statistical significance tests and measures of association were not included.  However, as the book points out, they can all be seen here.

From time to time changes on income distribution may be triggered by new governments with different philosophies or economic theories.  This happened in a number of countries when monetarism and neo-liberal ideas became common during the 1980s.  Legislation was introduced to weaken trade union powers and changes in taxes and benefits were introduced which contributed to a widening in income differences.  However, although it can be said that a change in political and economic ideology contributed to widening income differences, there is no doubt that governments did not intend to weaken community life or to increase levels of violence, teenage birth rates, drug abuse or any of the other problems which go with greater inequality.  These were all unintended consequences of widening income differences. 

Nor can an increase in health and social problems be the cause of widening income differences.  As we have seen, a wide range of health and social problems tend to move together – countries which do badly (or well) on one outcome tend to do badly (or well) on others. If they were not all results of inequality but were instead separate causes of inequality, that would not explain why they move together.  Indeed it is not plausible to think that problems such as homicide, obesity and low standards of child wellbeing – which are all associated with inequality – could be a cause of it.

No. On p. 80 of The Spirit Level we show that there is no correlation between life expectancy and total medical expenditure per head in different countries.  Although medical care is important for many aspects of the quality of life, such as hip and knee replacements, cataract and hernia operations, it looks as if the huge differences in the rates at which people get life threatening conditions such as cancer or heart disease, overshadow the differences made to survival by the quality of medical care.  The fact that the vast majority of medical expenditure on each person is spent during their last year of life suggests that its ability to extend the length of life is limited. 

OECD figures on government social expenditure as a proportion of national income are, like medical expenditure, unrelated to the extent of problems as measured by the Index of Health and Social Problems (see p.177 of The Spirit Level).  Many services may be seen as attempts to cope with problems created elsewhere in society.  In an important sense they exist to pick up the pieces but are rarely the main determinants of the scale of problems which exist in each society.  As well as being true of medical care, criminology studies suggest that differences in policing do not exert a major influence on levels of crime.  Even in education, it is well known that early experience, family life and socioeconomic circumstances exert overwhelmingly powerful influences on ‘school readiness’ and subsequent levels of educational achievement.

No. If you look at the graph in which all the separate outcomes are combined into one Index of Health and Social Problems (Figure 2.2 in The Spirit Level),  you will see that even if all the English speaking countries were excluded, there is still a powerful relationship between inequality and the index among the remaining countries.  The same is true of the UNICEF Index of Child Wellbeing.  A statistically significant relationship also remains between inequality and the Index of Health and Social Problems if you instead remove all the Nordic countries.  Indeed, the relationship is still statistically significant if countries at each end of the distribution – Sweden, Norway, Finland, USA and UK – are removed.  The association between inequality and social dysfunction cannot be explained simply by cultural differences.

Although some of the more equal countries – such as Norway, Sweden and Finland – have small populations and perform well, there are also small countries like Singapore and Portugal which do particularly badly. And the two countries in our data with the largest populations are the USA and Japan which lie at opposite ends of the inequality spectrum. If you control for population size, the correlations measuring how closely related outcomes are to inequality become even stronger. 

Some people wonder whether measures of income inequality are actually biased by population size – making countries with large populations appear more unequal. However, measures of income inequality are all designed to be unaffected by population size. 

People sometimes wonder whether the countries which do well are more homogenous and have fewer ethnic divisions than ones which do badly.  There are several points to keep in mind when considering this.  

First, an international study, which collected data on the ethnic mix in each country found that it did not explain the association between health and equality.  Interestingly, a very similar proportion of the population of Sweden and the USA are foreign born, and Spain is more equal and does better than its neighbour, Portugal, despite having a larger migrant population.

Among the US states income distribution tends to be more unequal in states in which a higher proportion of the population are African-American. One paper suggested this explained the relation between inequality and health.  Since then other papers have been published showing this is not so and a computational error was found in the paper which first made that suggestion.  In addition, in the more unequal states health is worse among both the black and white populations.  But migrant groups sometimes have unexpectedly good health.  In the USA, the largest group of migrants are Hispanic, predominantly from Mexico.  Although their levels of education and income are much like those of African Americans, for most outcomes their health is as good as that of the non-Hispanic white population.  That they do not seem to suffer the effects of their low social status is sometimes referred to as the ‘Hispanic paradox’.

However, insofar as ethnic divisions are related to inequality and may contribute to the effects of inequality, it is not of course skin colour itself – or for that matter religious or linguistic differences in themselves – which affect health.  Instead, they become important when they serve as markers of social status attracting stigmatisation, prejudice and discrimination. This means that rather than ethnic divisions involving quite separate processes from those through which inequality has its effects, they involve very much the same processes.  Whether the markers of social status differences are attributes of class alone or whether they include issues of language, religion or ethnicity, the underlying processes are basically the same. 

There has been rather little research looking at time trends and the few that there have been are confined to health.  However, there have now been something like 200 mainly cross-sectional studies of the relation between income distribution and health at different points in time in different settings.  It is of course impossible that the two would appear related at successive points in time unless they changed together over time.  If two things almost always appear together, then it is safe to assume that they move together.  (The possibility that they remain in the same place can be discounted: we know that there have been very substantial changes in income distribution.) 

The difficulty in tracking the links is the likelihood that long and differing lag periods are involved.  We know for instance that health in middle and old age is strongly influenced by early life experience – as well as experience over the intervening years. Each age group would require different lag periods, and different causes of death take different amounts of time to develop. On top of that, comparable international data on income distribution has not been available for very long.  In this situation it is impressive that studies have found relationships between changes in income inequality over time and changes in death rates. 

Rather than stimulating innovation and progress, great inequality wastes the talents of a large proportion of the population.  The evidence shows that it reduces children’s educational performance as well as reducing social mobility.  Economic studies of the relationship between the extent of inequality and economic growth rates have mixed results: most suggest that greater equality is beneficial to growth but a few suggest the opposite. As a check on how inequality might affect creativeness and innovation, we have now looked at the relationship between inequality and the number of patents granted per head of population. There is no tendency whatsoever for more unequal societies to gain more patent's per head than more equal ones. (N.B. This is a correction. Using published patents data that turned out to be inaccurate, we had previously stated that more equal societies actually had significantly higher patents per head).

Research on relative deprivation has found that if you ask people who they compare themselves with, they usually say it is people like themselves – such as neighbours, friends or relations. People sometimes suggest that income inequality must work through social comparisons, through people feeling out-done by neighbours who perhaps have a better car or bigger house. If so, it ought to have its most powerful effects when people getting very different incomes live close to each other and encounter each other face to face.

This kind of reasoning led some researchers to compare inequality and health measured in small areas. But a review of nearly 170 studies found that ones where inequality was measured in small areas were least likely to find a relationship between inequality and health.  The explanation is of course that small deprived neighbourhoods do not have poor health because of the inequality within them. They have bad health because they are deprived in relation to the wider society; and to capture that, inequality must be measured across society as a whole.  It is when inequality is measured across whole societies that it is most consistently related to outcomes such as health and homicide.

The explanation seems to be that what matters most is the scale of social differentiation across the whole society.  Social hierarchy is usually regarded as a national pyramid with small numbers of the rich at the top and the bulk of the population further down.  Rather than thinking of the effects of income distribution and of social class as separate influences on health and social functioning, it is more accurate to think of income inequality as providing the framework round which social class differentiation takes place.  Wider income differences lead to bigger social distances, more marked differentiation in terms of housing, cars, clothing and all the cultural markers of status, and so a more divided society.  In two books Robert Frank has shown how much people use income to express social status – see Frank RH. Luxury Fever: Why money fails to satisfy in an era of success. Free Press, N.Y. 1999; Frank RH. Falling Behind: How Rising Inequality Harms the Middle Class. University of California Press, Berkeley. 2007

The evidence suggests that the wellbeing of the poor is more closely tied to income levels than is the wellbeing of the better off. When parking fines or a washing machine repair make a major difference to a whole month’s budgeting and sometimes  lead people to borrow from loan sharks or fall behind with the rent, there can be little doubt that higher incomes would decrease the sense of desperation and hopelessness that are particularly common among the least well off.  

Not having a sense of being in control of things has been shown to be an important source of stress and harmful to health.  Inadequate incomes mean people face one crisis after another. As a result, family life suffers as financial issues become a frequent source of conflict.

There is a tendency for women’s status to be better in more equal countries, such as the Scandinavian countries, as well as in the more equal US states.  Japan is however an important exception.  Although Sweden and Japan are among the most equal of the rich developed societies and both do well in terms of health and a wide range of other outcomes, the position of women in these two societies is extraordinarily different.  

Where the overall income differences in society are bigger, women seem to suffer a bigger income disadvantage compared to men. But despite that, inequality sometimes seems to affect male health even more strongly than women’s health.  It looks as if male culture in more unequal societies is more macho, with more deaths associated with risk-taking behaviour and violence.  

In studies of health and development in poorer countries, women’s status (usually measured by comparing men and women’s educational levels) is associated with lower death rates for men, women and infants. Remarkably, in some studies lower status for women seems even worse for men’s than women’s health.  If women’s status is low, it may sometimes be an indication of more macho cultures which rebound particularly on male health.  Improvements in women’s status may depend on the development of a gentler more sociable society, less dominated by a more macho culture. 

No. The scale of income inequality is very different in different countries – and even in different states of the USA.  The more unequal of the rich developed countries (Singapore, USA, Portugal and the UK) are about twice as unequal as countries like Japan, Norway, Sweden and Finland.  There are also big differences in the levels of inequality in poorer countries.  Not only does the extent of inequality vary from country to country, it also varies over time.  Income inequality grew rapidly in Britain, particularly from the mid 1980s to the early 1990s.  The USA had an almost continuous increase in inequality from the mid 1970s to the early 1990s.  Graphs showing the trends can be found in the last chapter of  The Spirit Level.  Although income differences widened in many countries, they did not do so in all countries.  In the 20 years from the mid 1980s onwards, countries such as France, Belgium, Spain and Greece were among those which managed to avoid increasing inequality.

Governments in all rich countries control close to 40 percent of economic activity: they cannot avoid affecting income differences.  The increase in inequality in both Britain and the USA almost certainly reflects the neo-liberal economic policies of the governments in power. Differences in “market incomes” – that is income before taxes and benefits – can be reduced by strong trade unions, by minimum pay policies, by employee representatives on the board, by a public ethic intolerant of the “bonus culture” and so on.  They can also be reduced by taxes and benefits, particularly if more stringent action is taken to prevent tax avoidance. Other less direct influence on income differences include education policies and the management of the national economy. 

There are lots of different ways of measuring income distribution (see our Notes on Statistical Sources [external link]).  In The Spirit Level we averaged the figures published in the UN Human Development Reports for the years 2003 to 2006. The UN figures for the rich countries are basically the same as the ones published by the World Bank and came originally from the Luxembourg Income Study  (LIS) which was set up to help produce internationally comparable figures of income inequality.

Whether you use the OECD inequality figures for the mid 2000s or those from the UN makes rather little difference to the results. The OECD Gini coefficients of inequality are strongly correlated with our Index of Health and Social Problems (r=0.7, p<0.001) – only a little weaker than the relation illustrated in Figure 2.2 in The Spirit Level. The relationship is slightly weaker because Japan becomes an outlier on the OECD figures. Although we have a rule that we take the data exactly as given in our sources without picking and choosing, if you preferred to take Japan out of the analysis – on the grounds that there is so much difference in its inequality figures from these two sources – the relationship using the OECD figures becomes stronger still.

We do not know why the Gini coefficients published by the OECD and the UN are so different, they differ substantially for several countries, including Japan, the Netherlands, Belgium, Switzerland and France. We know of no good reason to prefer one data series over another.

 

However, we have checked the robustness of the analyses in The Spirit Level, using the OECD Gini coefficient for the mid-2000s.

 

The following are significantly associated (in the same directions as in The Spirit Level) with income inequality:

Health or social problem

Correlation coefficient (r)

Probability value (p)

Trust

-0.66

0.03

Women’s status*

-0.58

<0.01

Teenage births

0.64

<0.01

Social mobility

0.83

<0.01

Obesity - women

0.47

0.03

Child wellbeing

0.68

<0.01

Homicides

0.44

0.04

Imprisonment

0.51

0.02

Educational scores*

-0.46

0.046

Foreign aid % of GNI*

-0.80

<0.01

Global peace index

0.48

0.03

Child overweight*

0.74

<0.01

Infant mortality*

0.54

0.01

Low birth weight

0.55

0.01

Child mortality

0.51

0.02

Child conflict

0.54

0.02

Suicide

-0.52

0.01

Recycling

0.70

0.02

Patents per capita

-0.50

0.02

Paid maternity leave

-0.55

<0.01

Advertising

0.53

0.01

Police

0.52

0.04

Social expenditure

-0.6

<0.01

* stronger correlation than in The Spirit Level

 

The following are not statistically significantly associated when using the OECD measure:

 

Health or social problem

Correlation coefficient (r)

Probability value (p)

Life expectancy

-0.27

0.24

Obesity men and women combined

0.41

0.06

Mental health**

0.32

0.30

Calories

0.29

0.20

Drugs index***

0.38

0.09

Public expenditure on health

-0.26

0.30

** Close to significance for explaining differences in mental health among 5 Anglo-speaking nations (r=0.53, p=0.08)

 

*** Significant for opiates and close to significance for cocaine

 

The latest figures on income inequality from the UN Human Development Report are in the 2009 report. They no longer report the 20:20 ratio used in The Spirit Level, only the 10:10 ratio and the Gini coefficient.


Looking at the health and social outcomes that were not significantly related to the OECD inequality measure, all are significant using the 10:10 ratio and all but one using the UN Gini coefficient:

 

 

10:10 Ratio

Gini

Health or social problem

Correlation

(r)

Probability

(p)

Correlation (r)

Probability (p)

Life expectancy

-0.44

0.03

-0.33

0.12

Obesity men and women combined

0.54

0.01

0.47

0.03

Mental health**

0.62

0.02

0.54

0.05

Calories

0.42

0.06

0.49

0.02

Drugs index***

0.64

<0.01

0.58

<0.01

Public expenditure on health

-0.47

0.02

-0.49

0.02