Income and Health 3 How income affects health

Not only the poor

Although the evidence suggests that making incomes more equal would have its greatest effect on the health of those in relative poverty, it looks as if its impact on health is too great to be explained wholly by health improvements among the poor. In the early 1980s, when income differentials were not as wide as they are now, British life expectancy at birth was around three years shorter than that of Japan and the most egalitarian European societies.  About two years of this short-fall seems to be associated with differences in income distribution.

Unfortunately we do not know the life expectancy of the twelve percent of the British population who in 1981 came within the European Community definition of relative poverty (living on less than fifty percent of the average). Social class V, whose death rates we do know accounted for only about six percent of the economically active population and may therefore represent a more extreme group. However, let us assume that the twelve percent then living in relative poverty had death rates equivalent to those found in social class V. If we then calculate the effect on the population’s average life expectancy of bringing those death rates down to the average, we find such a change would account for only about six months of the two-year short fall in British life expectancy. To account for the whole two years it would be necessary to assume that the least well off quarter of the population overcame a health disadvantage twice as great as that suffered by social class V. Despite using many different measures of socio-economic disadvantage, no one has yet identified such a large proportion of the population at such a high risk.

If we were nevertheless to assume that all the improvement in life expectancy came from a levelling up process of this kind, it might be expected to lead to the disappearance of existing health inequalities. But the more egalitarian societies like Norway, Sweden and the Netherlands, have not abolished their internal health inequalities.  Though apparently smaller than the differences found in Britain, they still have substantial health inequalities.” While Japanese income distribution and life expectancy have improved rapidly, health gains have (in contrast to the British experience) been faster among blue-collar than white-collar workers but important differences still remain between them.

Together, these considerations suggest that the health differences between the more and less egalitarian developed countries cannot be attributed wholly to differences in the numbers and health of the relatively poor in each country. The implication is that the health of other sections of the population must also benefit from income redistribution.  If you take figures of the proportion of income going to the least well-off ten, twenty, thirty percent and so on in each country, you find that the relationship with life expectancy reaches its strongest when you take the proportion of income going to the bottom sixty or seventy percent of the population. This not only makes sense in terms of the size of the health benefits, but it rather neatly points to average income as the line which separates the gainers from the remainder. Because the distribution of income is skewed by a small number of very rich people who push the average up, a little over sixty percent of the population live on less than the average. (More sophisticated measures of inequality, such as the Gini coefficient, which measure inequality across the whole population are almost equally closely related to life expectancy.)

What are the links?

Greater income equality seems to create a healthier society as a whole. How can this be explained?  What sort of factors link health and income distribution? The first thing to remember is that we are dealing primarily – as figure 2 shows – with the effects of relative income and poverty, not with the effects of absolute income levels. It is not how rich or poor people are in absolute terms which matters, but how rich or poor they are in relation to others in their society.

Although some developed countries have per capita incomes three times as high as others there are no consistent differences in health between them. If it was absolute poverty which affected health we would expect health to improve as increasing wealth led to a diminution in absolute poverty. Levels of absolute poverty could only remain untouched by increasing affluence if income differentials tended to widen as countries got richer. However, there is no evidence that they do – rather the reverse.

Large differences in absolute income appear then to have little or no effect on mortality, but small differences in income distribution appear to have a large effect. This provides fairly clear evidence that it is relative standards, or differences, rather than absolute standards which matter.

If the issue is not absolute income levels, this suggests that health is no longer determined primarily by the directly physiological effects of the material circumstances in which people live. The importance of relative income implies that the crucial issue is what a person’s income or standard of living means in the social context of their society. The problem lies less in the physical consequences of the material conditions of life than with their psychological and emotional consequences.

An example may clarify the issue. There are still numerous ways in which differences in people’s physical circumstances continue to have a direct effect on their health. Housing is a case in point. Research has reliably identified a connection between high rates of some respiratory illnesses and the mould spores to which the occupants of damp housing are exposed.” Similarly, people living in poorer housing are less likely to have central heating and so are more likely to use types of heating which carry higher fire risks. There are many other examples of ways in which the least well off are more often exposed to direct physical health risks whose consequences they will suffer regardless of their state of mind. Tackling such factors must remain a high priority for public health policy.

The overwhelming importance of relative income suggests that physical exposure to material hazards such as poor housing are no longer the main determinants of health.  Much the largest part of the problem is not the material concomitants or consequences of relative deprivation, but the fact of relative deprivation itself. This means that we have to address ourselves to the more fundamental but more intangible problems of the psychological and social implications of income differences, of relative poverty and of having to live in conditions which are recognisably substandard – regardless of what affluence may have done to the standard.

A study which quite incidentally showed the primacy of the psychological over the material links between health and socio-economic disadvantage was a study of people made redundant when a factory closed. GPs found that increases in both major and minor illnesses among their patients dated from when redundancies were first announced – before people actually became unemployed.

Having said that, it may be that some of the health effects of relative poverty can be avoided by those few who have to some extent chosen a cheaper lifestyle. But for most people it is hard to live on a low income without financial worries and stress, without it cramping your style or limiting your social contacts and confidence, and without a sense of diminishing self-esteem and worthlessness.  Clearly, the effects of poverty on monks, political activists, self-sufficiency enthusiasts or committed artists may be very different from the effects on someone who would like nothing better than a well paid job. But in societies where appearances count for so much, few will have the emotional resources or the alternative sources of self-esteem to avoid the demeaning effects of a low income.

There are other effects which are unavoidable. The infrastructure of modern societies is constantly developing to meet the needs of people equipped with average or above average resources. Without the necessary resources to keep up, many daily activities are made increasingly difficult. For instance, not having a washing machine is not much of an inconvenience where everyone goes to a local launderette. But in a society where most people have washing machines and launderettes are rare, the poor may be reduced to washing clothes by hand in a council flat which, though provided with plumbing for a washing machine, may have only an unsuitably small sink. Similarly, where most people have a car, the tendency is to site amenities to suit car owners and then to cut back on public transport.

Unless people can afford to equip themselves with the normal range of consumer goods such as a washing machine, car, telephone, cooker and fridge, ordinary tasks become more time consuming and awkward. So much so, that the poor are increasingly excluded from participating in the ordinary life of society. The problem is inherently a relative one in which standards constantly change: watches became a necessity as timekeeping became more important many years ago. As shopping and home life become geared up to the use of frozen foods, it becomes harder to do without freezers. Similarly, as credit cards and credit-worthiness play an expanding role in a wide variety of financial transactions, their lack becomes inconvenient and stigmatising.

At this level poverty is literally disabling.  And, like many forms of disability, poverty is also socially handicapping. People with inadequate incomes constantly find themselves at a social disadvantage and may often be unable to participate in ordinary social activities and maintain ordinary social relations. They may not be able to afford to go out with friends or pay for their children to go on a school trip. The expense of visiting family and friends may lead to less frequent contact. It becomes harder to entertain people or to fulfil some of the obligations that a good friend or relative might be expected to. Neediness and an inability to reciprocate turns relationships into one-sided dependencies which many friendships will not survive. As old friendships break up and the social activities needed to make new ones are out of reach, it is not surprising that surveys show that the poor are socially isolated.”

Even though, in terms of income levels, we are dealing with relative rather than absolute poverty, that does not mean that people will always give highest priority to the purchase of so-called ‘basic necessities’.  Having a drink with friends or buying your children clothes that you are not ashamed for them to be seen in may sometimes seem more important than healthy food or adequate heating.  Thus, incomes which are theoretically adequate to pay for ‘essentials’ may, in practice, not be enough to prevent people suffering some of the physiological effects of absolute poverty. This might explain why small surveys have found that a quarter of all parents on what was then Supplementary Benefit had only one meal a day and that at the end of the week a quarter of all people on unemployment benefit did not have enough money left for food. No doubt some of the seven percent of pensioners which a DHSS survey found were malnourished preferred to spend less on food in order to have a television to relieve the endless hours of boredom and isolation. Similarly it is clear that for some teenagers it is more important to extricate themselves from family conflicts than to have a roof over their heads.  As human beings, our social needs are often as pressing as our physiological needs, and minimum income levels must recognise this.

The state of mind and levels of stress induced by poverty can be seen most clearly in the results of a study of families in bed-and-breakfast accommodation.  Forty four percent of the women said they were unhappy most of the time, forty one percent were tired most of the time, thirty five percent often lost their temper, thirty four percent could not sleep at night, thirty three percent said their children got on top of them and twenty four percent said they burst into tears for no reason. Inadequate cooking facilities meant that many families could not even cook regular meals for themselves.

Although people in bed-and-breakfast accommodation may be only a tiny minority, they and others classified as homeless totalled almost 700,000 in 1989. The experience of coping on inadequate incomes is the lot of millions of other people on whom it must have a similar, if less extreme, psychological impact.

The health effects of some of these aspects of poverty and inequality are only too obvious. To point out that we are a long way away from the sense of happiness and well-being which have been found to be predictive of longevity may risk trivialising the picture. However it is worth emphasising that the health disadvantage of the poor is a double injustice: rather than being ‘short and sweet, life is short where its quality is poor. Health is often an indicator of the real quality of human life.

No doubt partly as a result of the lack of friends, lack of social support, additional stress and the undermining of self-esteem, the poor tend to smoke more, and the number of heavy drinkers and users of prescribed and unprescribed drugs among them may also increase.

Social relations

The key element in understanding the health effects not only of poverty but also of income differentials more generally is likely to be the combination of stress, insecurity and poor social relations. Factors such as stress, ‘social support, ‘confiding relationships’, social participation and self-esteem have all been shown to be closely related to health. Indeed, good social support is almost certainly protective against some of the health effects of stress, so have more stress and less social support are particularly vulnerable. The epidemiological evidence suggests that the whole range of social relations have an influence on health – from the most personal domestic relations to people’s participation in local community activities – including social relations in the workplace.  At the most personal level, surveys have found that women in lower classes are less likely to have the benefit of ‘confiding relationships. No doubt disputes over money, which are a common source of stress and marital conflict, contribute to this pattern. At other levels, a study of some 17,000 office based civil servants found that the most senior staff were not only more likely than junior staff to see a confidante daily, but were also much more likely to have social contact with neighbours and with people from work.

While psycho-social problems of stress, self-esteem and social relations are most severe among the poor, their effects will be felt with diminishing frequency all the way up the social scale. If the health disadvantage of the least well off is mainly a problem of relative position, of expectations and standards informed – as they must be – by comparison with others, then it is easy to see why the health effects of income distribution are not confined to the poor.  A sense of relative deprivation can exist in varying degrees over most of the income range – from the poorest right up to within striking distance of the richest.  Large income differentials have always been regarded as socially divisive: and close friendships are difficult between people with very different resources at their disposal.  It is likely that income differentials affect levels of stress, insecurity and the quality of social relations throughout society so that the more divided a society is, the more strain it is likely to place on individuals within it.

One might speculate that the war-time improvements in civilian life expectancy were partly a product of the sense of camaraderie and shared purpose which older people so frequently recall. Indeed, as well as an improved income distribution giving rise to better social relations, the sense of a common bond probably also facilitated the development of war-time policies intended to ensure minimum standards of provision for all.  It is hard to know here which caused which, and there are clearly possibilities for the development of virtuous – or, conversely, vicious – circles in this field. No doubt the growing sense of social cohesion during the war was an important factor contributing to the election of the 1945 Labour government committed to the expansion of the welfare state.

Societies with wide income differentials and inadequate anti-poverty policies, with high rates of homelessness and unemployment, might be expected to suffer from more crime, more street violence and more frequent use of illegal drugs. As well as the direct health effects of these factors, they are likely to lead to a further deterioration in social relations. There will be a rise in tension on the streets, an increased fear of strangers, and a growing concern for security and personal safety. The elderly and others who feel vulnerable will restrict their activities, avoid going out at night and perhaps even cease answering the door bell after dark, if they live alone. Although there has been no research on it, it is possible that people’s increased wariness, sense of insecurity and tension may be a widespread health hazard. Whether people feel insecure or fearful only in public places or also when at home alone, if social relations deteriorate to the point where they are predominantly a source of stress, rather than a support in dealing with it, that must have important implications for health.

We have tried to suggest some of the possible ways in which income distribution might exert its influence on health. There are of course numerous other possible pathways. Fortunately establishing causation does not, as people often suggest, mean showing mechanism. In the field of health much the strongest evidence of causation comes from randomised controlled trials which say nothing about mechanism. Thus we knew that aspirin reduced the incidence of heart attacks before anyone knew why, and even where randomised controlled trials were impossible, we knew that smoking caused lung cancer without knowing how. The statistical evidence on the relationship between income distribution and health is already strong and research on it is continuing. The evidence is however already sufficiently clear for it to be regarded, in analogous medical circumstances, as unethical to withhold treatment.

If we knew exactly how income redistribution benefited health it would be a less urgent addition to public health policy than it now is. If its influence could be accounted for wholly in terms of its impact on better known factors such as smoking and diet, it would merely provide an additional way of influencing causes of death over which people already have at least some control.  But in fact income redistribution is particularly important because it provides a way of reducing death rates even from causes of death over which we would otherwise have no control. With the exception of lung cancer, preventable risk factors account for comparatively little of most of the major causes of death.  Until our knowledge of the aetiology of the degenerative diseases is much better understood it will remain impossible to account for much of the influence income distribution has on death rates. In the meantime we would do well to remember that rather than detracting from the importance we should attach to it, it increases our need to rely on such factors. That so many causes of death are sensitive to socio-economic differences may be the most useful thing we know about them.