Inequality of income, illiteracy and medical care as determinants of infant mortality in underdeveloped countries

Abstract The primary purpose of the paper is to explain the differences in infant mortality rates for a sample of 46 underdeveloped countries. The study differs from earlier research in this field in the following respects. First, an attempt is made to include explanatory variables which have some bearing upon policy discussions. Secondly, considerable attention is devoted to the correct specification of the regression equation. Thirdly, an attempt is made to assess the sensitivity of the estimates to changes in (i) the composition of the sample and (ii) the observations on infant mortality. Finally, the data employed in the study appear to be somewhat less inaccurate than the figures which have been used elsewhere. The most satisfactory empirical results are obtained from a logarithmic model, in which the illiteracy rate for women (I), nurses per head of population (N), physicians per head (P), and the coefficient of variation (V) — a measure of inequality of incomes — appear as explanatory variables. The results for this model indicate that 73 per cent of the inter-country variation in the logarithm of the infant mortality rate (log D) can be explained by differences in log I, log N, log P and log V. It is worth emphasizing that the estimated parameters of this model are found to be fairly precise and robust, in contrast to the results obtained in earlier studies. An important finding which emerges from the econometric analysis is that the stage of economic development per se appears not to have a direct impact upon infant mortality. Another finding of considerable interest is that the results fail to uphold the hypothesis that high fertility is a cause ofhigh infant mortality It is also worth mentioning that infant mortality appears to be more sensitive to variations in the share of the exceptionally rich than to changes in the share of the comparatively poor. An examination of the policy implications of the model reveals that D may be expected to be very sensitive to changes in I, P and V. It is concluded that a rapid fall in the infant mortality rates of underdeveloped countries will not be achieved unless greater emphasis is placed upon reducing inequality of incomes and augmenting the social infrastructure. Additional informationNotes on contributorsA. T. Flegg The author would like to thank Eduardo Arriaga, Derek Braddon, James Foreman-Peck, David Golby, Charles Harvey, Guy Judge, Alberto Palloni, Simon Price, Gerry Rodgers and Reg Ruel for their helpful suggestions; Dr Palloni for providing unpublished data on infant mortality; the U.S. Bureau of the Census for its assistance; and to his colleagues at Northridge for their critical appraisal of this study.