(see Poster Gallery) Investigating U.S. State-Level Income Inequality As a Determinant of Population Health: Theory, Evidence, and Directions Forward

Abstract

Background/Purpose: The income inequality (II) hypothesis (IIH) suggests that greater II negatively impacts population health. While a sizable IIH-specific academic literature has been exploring this phenomenon, this literature is generally not conversant with or integrative of either historical scholarly literature on II or modern IIH-peripheral literatures. In the IIH-specific literature, open questions remain regarding theoretical effect pathways that mediate the II-health relationship (i.e., individual, group, or policy effects) and what levels of analysis are most relevant (i.e., local, state, or national).

Methods: I undertake a review of historical, IIH-peripheral, and theoretical IIH-specific literatures assessing what the key considerations of these literatures have been and what insights about effect pathways and levels of analysis are available. I also review empirical IIH-specific literature at the state level to assess overall supportiveness of findings. Finally, I analyze new empirical data from two, independent, large-sample (n=101,782–759,262), nationally representative datasets (National Longitudinal Mortality Study and Behavioral Risk Factor Surveillance System), across a range of health outcomes (suicide-, homicide-, heart disease-, and total-mortality; general health; depression; anxiety; and heavy drinking), contrasting multiple theoretically based model formulations, with sensitivity checks for variable formulation and II measurement. Collectively, these present a diverse and strong range of new evidence regarding questions of theoretical pathway and level of analysis.

Results: Historical and peripheral literatures strongly affirm that the IIH examination is well-founded. Historical literature emphasizes considerations of justice and both literatures more strongly favor theoretical pathways in the domain of policy/politics. Individual level effects are also common considerations. Further, the level of the nation, especially historically, is far more widely considered than smaller regions such as states, or local levels. Contrastingly, the IIH-specific literature heavily considers the state-level more than other literatures and has little exploration of the policy/politics domain, instead favoring effects occurring via group or individual phenomena. Theory is underdeveloped and model construction inconsistencies abound. The empirical literature overall moderately supports the IIH at the state level. My empirical findings suggest that across 400 models, only slightly more than half (53%) support the IIH at the state level. These findings did not vary meaningfully with respect to time and held for the full range of outcomes, formulations, II-measures, and model specifications.

Conclusion/Implications: The new empirical findings presented support a weak or null interpretation of the IIH at the state level. Inconsistent supportiveness of prior empirical studies as well as theoretical misalignment with policy, individual, and group effects, indicate that state-level effects may be weaker overall and suggest why nation-level effects are likely to be stronger. Based on the historical and peripheral literature, IIH-specific research would benefit from considering perceived injustice as a potential mechanism in the production of diminished health. Finally, considering the historical and peripheral literature as well as IIH-specific suggestions, the near omission of the policy domain from IIH research should be reversed. We should not disregard compositional or poverty effects and should improve modeling and theory agreement across research which would more clearly entail falsification or support criteria for the IIH.