Titleholder Data and Estimation
The fitted model provides a good fit to titleholder data from 1955 to 2019. Maximum likelihood estimation is based on the titleholder transition likelihood. The integrals in the likelihood are evaluated numerically because the gamma-Gomper…
1 sources - 6 claims
The fitted model provides a good fit to titleholder data from 1955 to 2019. Maximum likelihood estimation is based on the titleholder transition likelihood. The integrals in the likelihood are evaluated numerically because the gamma-Gompertz parameter a depends on the integration variable. Optimization was implemented in Python with Nelder-Mead and repeated tolerance adjustment. The titleholder data are transformed into death time, deceased titleholder age, and successor age triples. The empirical application uses Gerontology Research Group world-oldest-person titleholder data since 1955.