Prior Work

The study challenges earlier estimates that suggested stagnation or losses in some states, especially Southern states and female populations. The paper presents its formula as more explicit than Green's main general result because it avoids computing extremal non-negative solutions. Dickinson's result covers only a subset of the diagonals addressed by this paper. Earlier Holford et al results likely reflected data and modelling limitations, according to this study. For k=1,m=2, the relevant Pascal diagonal sums give Fibonacci numbers. Green generalized Fibonacci diagonal sums to x+my=n, while Dickinson and Raab handled related or partial cases. Age-period-cohort models have an identification problem because period is linearly dependent on age and cohort. Holford et al used data beginning in 1969 to infer mortality experiences for cohorts born as early as 1900.