Forecasting Period and Cohort Mortality Trends

Christina Bohk, University of Rostock

Examining the pros and cons of period and cohort mortality has a long history in demography. While period measures analyze mortality of synthetic cohorts, cohort measures analyze mortality of real cohorts. Apart from these principal differences, there is a huge amount of literature comparing period and cohort life expectancy, discussing potential distortions due to changing age-specific death rates. To analyze the impact of period and cohort mortality on life expectancy, we extend our period model so that it can also forecast cohort mortality data. We then apply both versions of our model to Danish female mortality. Our results suggest that expected years of life are likely to further increase according to both measures, though cohort life expectancy is forecasted to remain less volatile and, on average, 8.3 years above period life expectancy. Moreover, forecast uncertainty appears to increase faster in the period than in the cohort scenario.