Relation between levels of infant‐, child‐ and maternal‐mortality and their rates of decline

Purpose - The purpose of this paper is to study the cross‐country relation between initial levels of infant‐, child‐ and maternal‐mortality and their rates of decline so as to see whether the so‐called Matthew effect or the inverse‐care principle operates relative to these three important health indicators. Design/methodology/approach - Data on the three variables for a large number of countries covering several periods between 1950 and 2007 are considered. Signs and significance of correlations between initial levels and the rates of decline over the period, and of coefficients of initial levels in regressions of rates of decline on the initial level, are studied. Findings - First, in a broad global context, higher initial levels of mortality are associated with significantly lower rates of decline in each of the three indicators for every period, thus providing strong support to the operation of the inverse‐care principle and the Matthew effect. Second, the high‐income countries (and transition economies) deviate from the global pattern. Third, following Hart's suggestion, the parametric contrast between the high income and the developing groups may be interpreted as indicative of stronger government intervention in the healthcare sector in high‐income countries. Fourth, the contrast may thus indicate the desirability of greater government intervention in provision of healthcare in developing countries. Fifth, operation of the inverse‐care principle and the Matthew effect is observed even in the absence of high‐HIV prevalence. Sixth, the observed negative covariation between initial mortality and its rate of decline implies cross‐country divergence in these core indicators of health. Originality/value - First, this is the only study to investigate the operation of the inverse‐care principle relative to infant mortality for such a large number of countries and such a long period. Second, it is also the only study to extend the investigation to child‐mortality and maternal‐mortality, which are heavily emphasized in the millennium development goals. Third, the patterns are studied not only merely for the entire set of countries, but also for several subgroups. Fourth, the observed parametric contrasts are interpreted as possibly reflecting the importance of government intervention in the healthcare sector in mitigating the operation of the inverse‐care phenomenon. Fifth, an effort is made to factor out the role of HIV so as to show that the pattern is not significantly altered by high prevalence of HIV in poor countries. Sixth, the implied cross‐country divergence in these important health variables is suggestive of the need for caution in interpreting the conclusions stated by some scholars about convergence in several quality‐of‐life indicators. Last, contrary to what some scholars have suggested, not merely does it not seem to be the case that the inverse‐care proposition relative to infant mortality is observed only in exceptional cases, but the reported evidence suggests that the proposition holds globally over long periods even for child‐ and maternal‐mortality.