Income distribution and health : the role of individual and municipality levels

Abstract

The question whether income inequality is detrimental for health is still, after almost a quarter of a century and an impressive amount of literature, debated. The purpose of this thesis is to explore the association between income inequality and different health outcomes in a relatively egalitarian country during the 1990s and early 2000s, thereby contributing to a better understanding of the multitude of interpretations. Specific aims are to analyse whether there is an association between income distribution in Swedish municipalities and risk of death and to test the hypothesis that manual workers are at higher risk of death than are non-manual employees when living in municipalities with higher income unequal. Further, to disentangle the effects of income inequality and residential segregation, measured as parish level homogeneity, on acute myocardial infarction and to test the hypothesis that income inequality at the municipality level, as a marker of social stratification within the municipalities, would be associated with risk of attempted suicide.

The data used for the four studies in the thesis were compiled as two data sets. The mid-agedata set comprised all people 40 - 64 years in 1990 Swedish census and linked with the national cause-of-death register and the National Patient Register, altogether 2.5 million individuals. The second data set consisted of all individuals living in Sweden at age 16 and born 1972 or 1977 (N= 213 395). The unique personal identification number used in Sweden enabled linkage between the national inpatient register at the National Board of Health and Welfare and Statistics Sweden.

The effect from income inequality on mortality was overestimated using an ordinary least square regression as compared to the estimates in a hierarchical analysis. The latter showed an inverse association with the relative risk which, once adjustment for individual level variables was made, disappeared. We used data comprising all people 40 - 64 years and being employed in 1990 to analyse the hypothesis that manual workers had a higher “vulnerability” than non-manual employees to income inequality. The results indicated that unskilled manual workers living in municipalities with higher income inequality had a higher mortality risk than their peers living in more income egalitarian municipalities. For high level non-manual employees the risk was inverse.

We studied how the association between AMI and income inequality at the municipality level changed when parish level homogeneity regarding affluence and disadvantage respectively, was included in the analysis. There was an overall inverse association between income inequality and AMI which seemed to be explained by the level of homogeneity regarding affluence at the parish level within the municipalities. We noted that the affluent parishes tended to be within the municipalities within the three largest metropolitan areas Stockholm, Gothenburg and Malmö. To better understand the role of economic activity we used the concept of economic regions in analyses of associations between income inequality, economic deprivation at parish and individual level and attempted suicide (AS) among youth. We found no overall association between income inequality and AS but once adjustment for economic region was made, there was a significant inverse association between income inequality and AS, strongest in the three metropolitan areas. Although there was a strong association between the degree of economic deprivation at the parish level and AS it did not explain much of the association between income inequality and attempted suicide at the municipality level.

The overall conclusion from the four studies is that there seems to be an inconsistent (weak) association between income inequality and health, dependent on which context analyses are made and the choice of outcome. However, the studies raise a number of issues related to the complexity in which the research question is embedded, for example the lack of data on welfare institutions or other potentially “buffering” mechanisms.