Socioeconomic determinants of health : Health and social cohesion: why care about income inequality?BMJ 1997; 314 doi: http://dx.doi.org/10.1136/bmj.314.7086.1037 (Published 05 April 1997) Cite this as: BMJ 1997;314:1037
- a Department of Health and Social Behavior, Harvard School of Public Health, 677 Huntington Avenue, Boston, MA 02115, USA
- b Department of Health Policy and Management, Harvard School of Public Health
- Correspondence to: Dr Kawachi
Throughout the world, wealth and income are becoming more concentrated. Growing evidence suggests that the distribution of income–in addition to the absolute standard of living enjoyed by the poor–is a key determinant of population health. A large gap between rich people and poor people leads to higher mortality through the breakdown of social cohesion. The recent surge in income inequality in many countries has been accompanied by a marked increase in the residential concentration of poverty and affluence. Residential segregation diminishes the opportunities for social cohesion. Income inequality has spillover effects on society at large, including increased rates of crime and violence, impeded productivity and economic growth, and the impaired functioning of representative democracy. The extent of inequality in society is often a consequence of explicit policies and public choice. Reducing income inequality offers the prospect of greater social cohesiveness and better population health.
Income inequality and mortality
The world's wealth is becoming more concentrated. According to the 1996 United Nations Human Development Report, the world's 358 richest individuals control economic assets equivalent to the combined annual incomes of the poor countries that are home to 45% of the world's population.1 In the past 20 years, many countries including the United States and Britain have experienced soaring rates of income inequality. Do these trends matter for the health of populations?
No one would dispute that poverty is bad for health. In general, the lower the material standard of living (as measured by indicators like income) the worse is the level of health, whether measured by mortality, morbidity, or quality of life. In the United States, which is supposedly the richest country in the world, poverty still accounts for nearly 6% of all adult mortality.2
Aside from the evidence on absolute deprivation, there is growing evidence that the relative distribution of income in a society matters in its own right for population health. This thesis, which has become most closely identified with the work of Richard Wilkinson,3 4 has been replicated in nearly a dozen studies internationally.4 Although some questions have been raised about the international evidence linking income inequality to mortality,5 three recent studies reported in this journal–two from the United States6 7 and one from Britain8–have suggested that income inequality predicts excess mortality within individual countries. In the American study by Kennedy et al, income inequality at the state level was strongly correlated with total mortality rates (r=0.54, P<0.05), even after median income, poverty rates, smoking prevalence, and race were taken into account.6 Income inequality was measured in that study by the Robin Hood index, which is the proportion of aggregate income that needs to be redistributed from the rich to the poor so as to achieve equality of incomes. A 1% rise in the Robin Hood Index was associated with an excess mortality of 21.7 deaths per 100 000 (95% confidence interval 6.6 to 36.7), suggesting that even a modest reduction in inequality could have an important impact on population health. The maldistribution of income was related not only to total mortality but also to infant mortality, homicides, and deaths from cardiovascular disease and neoplasms.
Social cohesion and income inequality
The repeated corroboration of the hypothesis that income inequality is harmful to health has spurred the search for the mechanisms underlying this relation. Some hypothesised pathways include psychologically harmful effects of relative deprivation and the lack of investment in human capital that is frequently evident in societies that tolerate large income differentials.9 It is also possible that some other exogenous factor, such as racial discrimination in the United States, accounts for both income inequality and excess mortality. Much work remains to be carried out in sorting through these possibilities.
One notion that has existed for some time is that a widening of the gap between the rich and poor might result in damage to the social fabric. In a seminal essay on the dysfunctions of social stratification published in 1953, Melvin Tumin speculated that “to the extent that inequalities in social rewards cannot be made fully acceptable to the less privileged in a society, social stratification systems function to encourage hostility, suspicion and distrust among the various segments of society and thus to limit the possibilities of extensive social integration.”10 In his latest book, Wilkinson provides case studies of societies that at certain points in history underwent either a rapid compression of the income distribution (Britain during the two world wars) or a rapid widening of income differentials (the Italian-American community of Roseto in Pennsylvania during the 1960s).4 In wartime Britain, narrowing of income differentials was accompanied by a greater sense of solidarity and social cohesion as well as dramatic improvements in life expectancy. In contrast, in the originally closeknit town of Roseto, rapid economic change in the 1960s opened the gap between rich people and poor people. The resulting breakdown of community cohesion was followed by a sharp increase in deaths from coronary disease.11
That social cohesion enhances wellbeing is by now a well established fact. Ever since Durkheim's study of the causes of suicide,12 numerous epidemiological studies have shown that people who are socially integrated live longer.13 14 15 Socially isolated people die at two to three times the rate of well connected people, presumably reflecting the former's limited access to sources of emotional support, instrumental support (for example, financial aid), and other forms of support. But what has been missing from recent epidemiological studies of social relationships and health is the social context in which people lead their lives. In other words, by focusing on the outcomes of socially isolated (or well connected) individuals, epidemiology has neglected the possibility that entire communities or societies might be lacking in social connections.
Fortunately, there has been a renaissance in the notion of community cohesiveness, with the publication in 1993 of a work by an American political scientist, Robert Putnam. In Making Democracy Work he sought to measure the strength of social cohesion–what he termed “social capital”–within regions of Italy.16 The purpose of his 20 year study was to attempt to explain the performance of local governments, which were introduced to Italy in 1970. Local government performance in each region of Italy was assessed by its responsiveness to constituents and its efficiency in conducting the public's business. According to Putnam, the stock of social capital in a region–for example, as measured by the density of citizens' participation in community organisations (choral societies, soccer leagues, Rotary clubs, and the like) turned out to be the best predictor of local government performance. Citizens living in regions characterised by high levels of social capital were more likely to trust their fellow citizens and to value solidarity, equality, and mutual tolerance. They were also blessed with highly functioning governments.
What can public health learn from this quasi-experiment in Italy? Although much work is needed in refining the notion of social capital, Kawachi et al recently carried out an analysis of income inequality in the United States and its relation to social capital, as defined by two of the indicators described by Putnam: levels of civic trust and density of associational membership. Data on social capital were obtained in 39 states from a survey conducted by the National Opinions Research Center between 1986 and 1990.17 The survey asked respondents in each state whether “Most people can be trusted–or would most people try to take advantage of you if they got the chance?” The percentage of citizens who thought that people try to take advantage (suggesting low levels of civic trust) was highly correlated with the degree of income inequality in each state (r=0.7, P<0.0001). Similarly, density of associational life, as gauged by the per capita membership of groups (church groups, sports groups, fraternal organizations, labour unions, and so on), was correlated with income inequality (r=−0.4, P<0.01). In turn, both the degree of civic distrust and paucity of associational life were strongly correlated with overall mortality (fig 2). The effect of income inequality on mortality thus seemed to be mediated through the withering of social capital.
The intervening effects of residential segregation
Concepts like income inequality and social capital are inherently “ecological”–that is, they are characteristics of places, not individuals. To understand the linkages between such variables, further research needs to focus on where people live rather than on the behaviours of individuals.18 Several researchers have attempted to separate out individual and area effects on mortality: is the health of people with any given level of individual socioeconomic characteristic better or worse according to whether they live in a rich or poor area?19 20 21 22 The assumption underlying such studies is that poor neighbourhoods are worse physical environments and lack amenities such as public transportation, access to primary care, banking facilities, and retail choice in healthy foods. Few of these studies have dealt with the impact of inequality itself. Indeed, not all studies have shown that poor people have worse health if they live in a poor area rather than in a rich one.20 This lack of unanimity may be the result of the sense of relative deprivation running counter to the effects of the wider environment: a poor person living in an affluent area may have a better environment but may also feel relatively poorer.
Medical demographers have an established tradition in studying the quantifiable characteristics of neighbourhoods. One such line of research–on the health effects of residential segregation–offers a promising approach by which to link the effects of income inequality on social disintegration.23 Accompanying the surge in income inequality in the United States since the mid-1970s, the spatial concentration of poverty has increased sharply. Between 1970 and 1990, the percentage of urban poor Americans living in non-poor neighbourhoods (defined as having poverty rates below 20%) declined from 45% to 31%, while the percentage living in poor neighbourhoods (poverty rates between 20% and 40%) increased from 38% to 41%. Meanwhile, the proportion living in very poor neighbourhoods (over 40% poverty) grew from 17% to 28%.24 Such patterns of residential concentration impose a double burden on poor people: not only do they have to grapple with the multiple problems arising from their own lack of income, they also have to deal with the social effects of living in a neighbourhood where most of their neighbours are also poor.25 Sociologist William Julius Wilson coined the term “concentration effects” to describe the cumulative disadvantages that are heaped on the residents of urban ghettos.25 Wilson and Wacquant found that residents in extremely poor neighbourhoods were also less likely to report the presence of regular sources of social support, including a marital partner and close friends (fig 3).26
The trends in the spatial concentration of poor people fail to address what has happened at the opposite end of the income spectrum. According to Massey and Denton, affluence in the United States is even more highly concentrated spatially than poverty.24 In 1970, the typical affluent American family–defined as having an income level at least four times the poverty rate–lived in a neighbourhood that was 39% affluent; by 1990, this had increased to 52%. In other words, the typical affluent person lived in a neighbourhood where more than half the residents were also rich. In contrast to people living in poverty stricken neighbourhoods, residents of affluent neighbourhoods benefit from better equipped public schools, higher quality public amenities, and more generous municipal services–all financed through higher property tax revenues. At the same time, the children of the privileged are more likely to socialise with other children of well educated and successful parents, thereby ensuring the social reproduction of material and cultural advantage.25 Without question, the privileged classes actively invest in their own social capital when they retreat to affluent residential enclaves. But the segregated nature of the resulting forms of social capital will tend to undermine social cohesion within society at large.
The social consequences of income inequality
Why should society really care about the extent of income inequality? Firstly, because income inequality induces “spillover” effects on quality of life, even for people not normally affected by material wants. Wide income disparities result in frustration, stress, and family disruption, which then increase the rates of crime, violence, and homicide.4 Those who can afford to will be increasingly forced to flee to walled compounds equipped with round the clock security systems, as has already happened in some American communities. Middle class flight from poor neighbourhoods results in the progressive deterioration of the public education system and the erosion of support for public schools. As Kaplan and others have shown, wide income disparities tend to coexist with underinvestment in human capital, measured in a variety of ways including high school drop out rates, reduced public spending on education, and lower literacy rates.7 The rise of an “underclass” of poorly educated and underskilled citizens means that society will ultimately pay the cost through low productivity and slow economic growth. Finally, as Putnam has suggested, the breakdown of social cohesion brought about by income inequality threatens the functioning of democracy. Low levels of civic trust spill over into lack of trust and confidence in government.16 To give an example from the United States, there is a strong correlation between lack of civic trust and low voter turnout at elections (fig 4). It is already known that the votes of the poor are underrepresented at election time. Political representation is further distorted by inequalities in political campaign donations across different income groups. In the United States it is estimated that the richest 3% of the voting population accounts for 35% of all private campaign donations during presidential elections.27
To a large extent it is a matter of public choice as to how much inequality a society should tolerate. The danger is that a society that becomes depleted of its stocks of social capital could enter into a vicious cycle–one in which lack of trust and civic engagement reinforces a kind of democracy in which public policy is no longer the outcome of collective deliberation about the public interest, but rather the residue of campaign strategy.28 The alternative is to put a halt to the growth in income inequality, which offers the hope of revitalising social capital at the same time as improving the health of the whole population.
Funding: IK and BPK are recipients of the Robert Wood Johnson Foundation Investigator Awards in Health Policy Research.
Conflict of interest: None.