Infant mortality has declined steadily in the U.S. since 1960 due largely to advances in medical technology and public health, and to the extension of medical coverage to the poor via Medicaid. But improvements in the U.S. have not kept pace with progress in other countries. In the 1960s, the U.S. had the 13th lowest rate of infant mortality in the world; but by 2005-10, the U.S. had fallen to 43rd place. The question is why. Arline Geronimus, John Bound, and Javier Rodriguez suggest that the answer may be partially a matter of politics. In a recent analysis of infant mortality data running back to 1960, they found that infant mortality nationwide is slightly but significantly lower than its long term trend line when Democratic presidents are in the White House, and slightly but significantly higher than trend during Republican presidencies. This is a striking, but mysterious result – one which begs for additional torture-testing to see whether it will hold up, and for additional probing to explore how this political effect might arise.
The project would begin by assembling a comprehensive new data set that will combine published data at the national, state, and county level on (1) infant mortality from the National Vital Statistical Report Series; (2) demographic data on the U.S. population from census; (3) other health related indicators (alcohol consumption, smoking, etc.) from the National Institutes of Health, the Centers for Disease Control and Prevention and other sources; and (4) political information from the Congressional Quarterly, the U.S. House of Representatives Congressional Directory, and the Congressional Budget office. County- and state-level data on infant mortality exists in published form back to 1959 and will be available in digital form from a colleague. The local level political data exist but are highly dispersed throughout institutional data banks. The same is true of the health data.
With this data set in hand, Geronimus and her colleagues propose to conduct an extensive set of analyses to see whether, and if so how, the political effect on infant mortality plays out at the state and local level. To begin, they would simply see whether the political effect holds up at lower levels of partisan control. Does the party in control of the governorship or the state legislature affect within-state infant mortality trends? They would also undertake extensive descriptive work to examine the incidence of the national political effect in different areas of the country. Geronimus and her colleagues also plan to examine the way in which the effectiveness of national health care policies are influenced by party control at different levels of government. They will look for interactions between federal, state, and local levels of partisan control – to see how national policies are mediated and moderated by local political control. They plan to do this by comparing the compound policy effects of different party-composition dyads or triads within and between jurisdictional levels. In addition, they intend to explore the way long-term effects of national health policies might explain the deviations from infant mortality trends observed under Democratic and Republican administrations. Prior research has demonstrated significant but slowly unfolding positive effects of national programs like Medicaid in reducing indicators of health inequality like infant mortality, albeit with substantial local variation in impact. Geronimus et al. plan to estimate these lagged effects and see how well they mesh with their own work on the influence of partisan control of the federal administration on infant mortality.