Mortality Concentration in the China Multi-Generational Panel Dataset, 1749-1909: Decomposition by Household and Community

James E. Oeppen, Max Planck Institute for Demographic Research

The analysis concerns mortality concentration within the China Multi-Generational Panel Dataset: using triennial observations on survival from 1749 to 1909. It uses a straightforward and simple approach to exactly decompose the observed change at the aggregate population level into relative contributions from average change in unit performance, change associated with withdrawal, and change caused by recruitment. This decomposition has implications for the understanding of mortality and the extent to which change is concentrated within families and communities; and how selection effects at the unit level contribute to changes in the population. Heterogeneity in mortality experience between aggregates such as households and communities means that the dynamics experienced by a population may not reflect the experience of the sub-units. This problem of possible compositional effects is compounded if the population is subject to recruitment and attrition of entities. Both of these effects may be observed in the Chinese data.