Cohort Analysis of Fertility

Abstract

IN I933 there were 2,3I2,000 births in the United States and the crude birth rate was i8.4. In I947 there were 3,876,ooo births and the crude birth rate was 27.0. In other words the number of births in I947 was 68 per cent larger than it was I4 years earlier, and the crude birth rate 47 per cent higher. Since I947 there have been only small declines. In I948 the number of births was approximately 3,7I5,ooo and the crude birth rate 25.4. Present indications are that in I949 the number of births will be about 3,750,000 and the crude birth rate 25.2. It is generally agreed that the much higher fertility of I947-49 than of I933 is due primarily to conditions associated with the great depression of the early I930's and with postwar demobilization and prosperity. Opinions differ widely, however, as to the extent to which the I94749 boom in babies represents (a) a reversal of the long-time downward trend in average family size, and (b) an unusual concentration in a few years of the births occurring with the continuance of the downward trend. The cohort material to be presented here In an attempt to evaluate the importance of these two explanations relates to native white women. The foreign-born contingent is omitted because it has decreased rapidly in recent years and now constitutes less than 5 per cent of the women in the reproductive ages.' The nonwhite group is omitted partly because it also is relatively small (about iI per cent) and partly because the data for it are substantially less reliable than those for native whites. But since native white women have composed over 75 per cent of all women aged I5-44 since I920 and since the relative rise in the number of births and the crude birth rate since I933 has been somewhat greater for native white women than for all women, the conclusions to be drawn from a discussion of this group should apply almost equally well to the total population. The results obtained by utilizing the methodology involved in computing gross and net reproduction rates would seem at first glance to show that the long-time decline in average number of children per family has been reversed. The conventional gross reproduction rate of native white women is I04 for I933 and I52 for I947. If male births are included in the computations, these figures are raised to 2I4 and 313, respectively. This means that in a hypothetical cohort of women living through the reproductive period under the agespecific birth rates of I933 there would be 2I4 births per ioo women reaching age 50. And in a hypothetical cohort exposed to the I947 rates there would be 313 births per ioo women-almost one and one half times as many as in I933. This type of analysis may be carried a step farther by subdividing births by order of birth and observing the extent to which the reproduction rate is composed of first births, second births, etc. In doing so, however, it is important also to subdivide women by parity. The conventional procedure ignores parity and gives impossible results for certain hypothetical cohorts, the most fantastic being the computed occurrence of I38 first births per ioo women in the hypothetical cohort living to age 50 under I947 conditions ! 2 Obviously there cannot be more than ioo first births per ioo women, even theoretically.3 Actually