Income Disparity and the Mississippi River
The study can be examined on two levels. First, if you take the study at face value and believe there were no omissions or errors, the study does not support the argument that disparity is more important than absolute poverty. Second, there are omissions and errors in the study that make it even less conclusive.
At Face Value the Harvard Study Does Not Support the Importance of Disparity Over Absolute Poverty
The authors begin by explaining why they looked at state-to-state comparisons as opposed to country-to-country comparisons. "Previous studies . . . have entailed comparisons across different countries, raising the question of comparability and completeness of income data." In other words, there are reasons to question the reliability of country-to-country comparisons. Remember this important qualification the next time differences between Northern Europe and the United States are compared and used as rhetorical ammunition by the Left.
Absolute income has long been related to age-adjusted mortality. The authors "adjust" mortality figures for absolute income, smoking and household size to eliminate these effects. After all these corrections, income disparity and mortality have r-squared coefficients ranging from 25% and 50%. Income disparity is correlated, but only weakly so, to residual mortality after eliminating the large effect of absolute income. Income disparity generally accounts for less than half the small residual variability.
Even if you believe that this remaining association is real and significant, the association with absolute poverty is greater. Consequently, if the choice of public policy reduces to reducing absolute poverty or income disparity, greater reductions in mortality result from increasing wealth than decreases in disparity. The Liberal emphasis on disparity is not supported.
The authors themselves were careful to suggest that "income disparity may be a proxy for other indicators." They urged that
"caution be exercised in the observed associations. Some states may have a higher proportion of sick people for reasons other than the hypothesis under investigation, and the less egalitarian distribution of income in such states merely reflects the reduced earning capacity of sick people, who are also at higher risk of dying."
In other words, it is possible that sickness causes both higher mortality and higher income inequality.
Finally, even if there is a valid statistical correlation between income inequality and mortality, it means, all other things being equal, reducing income inequality may reduce mortality. However, societies are complex systems and all things are never equal. If, for example, I show a
Latest Articles