After the mortality improvements of the Golden Cohort, how has the Baby Boom generation fared? Good news and bad news.

Marion Burkimsher, University of Lausanne

This study looks at the mortality trends of successive cohorts born 1930-1960. We calculated the risk of dying in adulthood using data from the Human Mortality Database and compared 17 western countries. Specifically we focus on male mortality between ages 30-53, although female mortality and other age spans exhibit similar patterns.

We found four distinct patterns for the cohorts born in these year bands:

  1. 1930-1938: universal pattern of steadily improving mortality for successive cohorts.
  2. 1939-1947: some countries have wide variations between cohorts (Austria, Finland). The Dutch famine of 1945 modestly increased the adult mortality likelihood of children born then. The most significant pattern seen in 11 out of the 17 countries was a sharp improvement for the 1946 cohort compared to those born before and after.
  3. 1948-1955: slow or stalling improvements in mortality across cohorts; decline in some countries (USA, Portugal, France, Denmark, Spain).
  4. 1956-1960: resumption of mortality improvement trends.

Previous work on Swedish mortality has found that children of lower birth order have a better mortality outlook than higher order children. With the limited data available on births by birth order (only the USA and Switzerland) we found evidence that the proportion of first births to all births may partly explain the trends seen in cohort mortality. During the 1930s fertility declined, and the proportion of first births increased, leading to mortality improvements across successive cohorts. 1946 saw a peak in first births and markedly lower adult mortality for that specific cohort. Subsequent Baby Boom children, however, tended to be (on average) higher order births - and consequently experienced less improvement in their mortality than for earlier cohorts.


Case and Deaton (2015) highlighted the rising mortality ofmid-life whites in the United States, a finding that has gained much media attention.Previously, the ongoing improvementsin mortality throughout their life of the Golden Cohort (those born in the 1930s)has been discussed (eg. Goldring et al 2011). Heatmaps, such as those included in the Goldring paperand the chapter on mortality improvements in Rau et al (2018) illustrate thedifference between period effects (which run vertically) and cohort effects(which are diagonals). Period effects on mortality are quick to spot, withobvious causes such as epidemics, wars or heat waves. However, less study hasbeen done on cohort trends, although recent work includes Murphy (2010).

            Thisstudy looks at the adult mortality of cohorts born 1930-1960. We compare 17countries with data in the Human Mortality Database: Western Europe, North Americaand Australasia. For each cohort at each age we calculate the age-specific deathrate (ASDR) for each year. For the denominator we added to the exposedpopulation deaths that had already occurred. We then cumulated the ASDRs forvarious adult age bands to determine the likelihood that a person of aparticular cohort died between certain ages (specifically 15-43, 15-53, 15-60,30-53). Separate analyses were completed for men and women. The trends observedin each of these analyses were similar although the levels varied. The datapresented below is for male mortality (likelihood of dying) between 30-53.

            TheBaby Boom generation is commonly defined as those born 1946-1964. With most datasets only going to 2013, then the 1960 cohort is the youngest we can include.

            Figure1 summarises the trends for four bands of cohorts.

A.    Golden cohorts, born 1930-1938. There wasimproving mortality with each successive cohort (downward sloping lines).

B.    War babies and immediately post-war, born1939-1947. For some countries there have been wild fluctuations between cohortsborn in different years of the war (e.g. Austria and Finland). The most prevalentpattern, however, seen in 11 of the countries, is the sharp improvement inmortality for the 1946 cohort compared to those born in the years immediatelyprior and after. The countries not affected by this were in Scandinavia (plottedwith dotted lines) and southern Europe (Spain and Portugal, dashed lines). A smallpeak in Dutch adult mortality of those born during the famine of 1945 can be seen,an effect that had been expected (Painter et al 2005).

C.    Early Baby-Boomers, born 1948-1955. The rate ofmortality improvement slowed markedly or stagnated. In some countries it evendeclined (USA, Portugal, France, Denmark, Spain).

D.   Later Baby Boomers, born 1956-1960. The downwardtrend of adult mortality resumed for these cohorts.

Figure 1: Likelihoodof dying between 30-53 by year of birth, male

            Whatmight explain the trends in cohort mortality? We have looked at one possibleexplanation, taking the cue from the sharp improvement for the 1946 cohort. Mightthe proportion of first births to all births be the decisive factor? Modin (2002) and Barclay and Kolk(2015) found that the lower the birth order the better the mortality prospects,first children faring best. Birth order data is available (only) for the USAand Switzerland and we plotted this against the mortality of males 30-53. Thetrends shown in Figure 2 suggest that birthorder proportions could explain some of the trends in adult mortality.

Figure 2. Adult malemortality and proportion of first births

References

Barclay, K. and Kolk, M., 2015. Birth order and mortality: apopulation-based cohort study. Demography, 52(2),pp.613-639.

Case, A. and Deaton, A., 2015. Rising morbidity and mortality in midlife amongwhite non-Hispanic Americans in the 21st century. Proceedingsof the National Academy of Sciences, 112(49), pp.15078-15083.

Goldring, S., Henretty, N., Mills, J., Johnson, K. andSmallwood, S., 2011. Mortality of the ÔGolden GenerationÕ: What can the ONSLongitudinal Study tell us? Population trends, 145(1),pp.203-232.

Kannisto, V., Christensen, K. and Vaupel, J.W., 1997. No increased mortality in later life for cohortsborn during famine. American Journal of Epidemiology, 145(11),pp.987-994.

Modin, B., 2002. Birth order and mortality: a life-long follow-up of 14,200boys and girls born in early 20th century Sweden. Socialscience & medicine, 54(7), pp.1051-1064.

Murphy, M., 2010. Reexamining the dominance of birth cohorteffects on mortality. Population and development review, 36(2),pp.365-390.

Painter, R.C., Roseboom, T.J., Bossuyt, P.M.,Osmond, C., Barker, D.J. and Bleker, O.P., 2005. Adult mortality at age 57 after prenatal exposureto the Dutch famine. European Journal of Epidemiology,20(8), pp.673-676.

Rau, R., Bohk-Ewald, C., Muszyńska,M.M. and Vaupel, J.W., 2018. Chapter 6: Surface Plotsof Rates of Mortality Improvement. In VisualizingMortality Dynamics in the Lexis Diagram (pp. 43-67). Springer.

Presented in Session 1196: Mortality and Longevity