The Geodemography of Infant Mortality in the Soviet Union, 1950-1990

Trends in infant mortality in the Soviet Union have been the subject of scientific, policy, and political interest. The reported increases in the IMR in the early 1970s drew special attention. Although they had only limited data available, analysts disa greed about the causes of this increase and whether it was real or an artifact of changes in data collection and data quality. We now have information on the IMRs for all 15 republics of the Soviet Union for almost every year from 1950 to 1990. Our anal ysis focuses on the geographic distribution of infant mortality and the relative contributions of regions to the all-Union rates. We do not adjust the reported IMRs for definitional differences or for underreporting.

Over time, the high and rising infant mortality in the traditionally Moslem republics had an increasing effect on infant mortality in the Soviet Union as a whole. However, atypical patterns of infant mortality by region and by urban-rural areas provide s trong evidence of serious underreporting of infant deaths, especially in rural areas, in the Moslem republics, and in earlier years. The patterns by region and urban-rural area became more plausible over time. It is unlikely that the rise in infant mor tality in the USSR in the early 1970s was "real." Infant mortality in the Soviet Union, especially in the Moslem republics, was high. However, the high level of infant mortality in the Soviet Union is not the same as a worsening trend. Because of chang es in data quality, it is not possible to say what the actual trend in the IMR was in the Moslem parts of the Soviet Union. This has implications for the interpretation of trends in IMR's in the post-Soviet period.

Data used: Official statistics on births and infant mortality rates for the USSR as a whole and for the 15 union republics for the years 1950-1990.