To-date the macroeconomic conditions-mortality literature on income-related inequality in mortality has relied on subgroup analysis, mainly using income as a stratification variable, but this nearly always causes selection bias yielding results that are hard to interpret. To solve this bad control problem, we apply a novel technique based on recentered influence function regression of overall income-related mortality measures, like the commonly used concentration index. We also highlight the importance of: i) measurement of relative versus absolute inequality; ii) measurement of inequality by population-level statistics of inequality (concentration indices) versus subgroup analysis; iii) measurement of short versus long-term income. We illustrate these issues and our suggested solution using detailed individual-level administrative data from Sweden. Our findings show that there overall is a (insignificant) counter-cyclical impact on mortality and its income-related inequality. During a sub-period of pronounced and significant counter-cyclical mortality we find support for accompanying counter-cyclical income-related inequality, but only when using short-term income.
|Published - 2020 nov. 9
|Lund University, Department of Economics