Social class mortality differentials, England and Wales

Stavan Shah and Nicholas Crafts, University of Warwick

[Project number 30062]

What has happened with respect to social class mortality differentials since 1921 is a "story which many are still in the process of telling" (Woods and Williams, 1995, p.106). My dissertation paper for the module Research in Applied Economics will study the patterns of mortality differentials in England and Wales. It also compares the subsets within social class mortality differentials, such as adults versus children, and male versus female is another interesting angle that has yet to be explored more explicitly. This module is a required assessment for Economics Department students at the University of Warwick who are in their final year.

Aaron Antonovsky (1967) analyzes long-term class differentials from before 1650 to after 1850. His three-period model - before 1650, 1650 to 1850, and since 1850 - is associated with, respectively, the absence, widening and then convergence of class mortality differentials.

Elsie Pamuk (1985) offers a different theory to the pattern of class mortality differentials. In her paper Social Class Inequality in Mortality from 1921 to 1972 in England and Wales she states that class inequality in mortality declined only in the 1920s. Her findings also show that inequality increased again during the 1950s and 1960s so that by the early 1970s it was greater than it had been before. Hattersley (1999) also states that by 1996 the differences between Social Class I and all other social classes were greater than earlier decades. Interestingly, inequality in mortality was increasing despite the introduction of the NHS in 1948.

Woods and Williams (1995) have also made a significant contribution to existing literature regarding adult and infant class mortality differentials. They are critiques of primarily Antonovsky's model and its foundations. Firstly, they suggest that even today there is always a 'trailing class' that lags behind when other classes experience more rapid decline in mortality. Secondly, they disagree with Antonovsky's statement that when mortality rates are extremely high or low, class differentials will tend to be small.

These papers are the background to my dissertation, and I will use Pamuk's (1985) methodology of data calculation, but I extend her work by three decades to 2001 as existing literature in this subject area is outdated. Pamuk (1985) uses the five-part social classification system, using an index based on the five classes and their proportionate share of population. Social classes are arranged from V to I - that is, from lowest to highest rank - on a horizontal axis and computing each class's range in the cumulative proportionate ranking of the population from lowest (0) to highest (1). To compute the age-standardized death rates for adults, age-specific rates for each class were applied. My paper will also have age-standardized death ratios as well as a slope index of inequality.

Top of page