A Flexible Two-Dimensional Mortality Model for Use in Indirect Estimation

John R. Wilmoth, University of California, Berkeley
Vladimir Canudas-Romo, Johns Hopkins University
Sarah Zureick, University of California, Berkeley
Cheryl C. Sawyer, United Nations

Mortality estimates for many populations are derived using model life tables. We propose a new model of age-specific mortality as a means of improving both the quality and the transparency of such estimates. We describe a flexible two-dimensional model of the age pattern of mortality and fit the model (separately by sex) to a collection of 616 period life tables from the Human Mortality Database. The fitted model is used to estimate full life tables given one or two pieces of information: child mortality only (5q0), or child mortality and adult mortality (5q0) and (45q15). Using empirical life tables from a variety of sources, we compare the performance of new and old methods. We find that the new model performs competitively against existing methods. If desired, the model can be applied in a flexible manner to incorporate non-quantitative information about the age pattern of mortality.