Abstract
Are claims of extraordinarily low mortality levels in the USSR justifiable? Applying a recently developed methodology appropriate for nonstable populations to 1959 and 1970 census data from the Soviet Union, we find that mortality is generally understated for the country as a whole and for various regions and republics. This is particularly so for the republics composing the Central Asian region and the Caucasus. Age overstatement appears to be extremely pronounced in the oldest segments of the population. Using the new methodology, we can derive the age distribution that is uniquely implied by a given life table and a set of age-specific rates of growth obtained from two censuses. When we use the official Soviet life tables in this procedure, we find that the reported number of centenarians is at least 28.9 percent overstated for males and 7.5 percent for females. If one were to posit that Soviet mortality during 1959 to 1970 was, in fact, no better than the Swedish mortality experience during roughly the same time period, then the true number of centenarians could be no more than 2 percent of that reported.
