This preview shows page 1. Sign up to view the full content.
Unformatted text preview: only the year of death and not allow for the time of death
Markov model allows for the time of death (preferred, if
complete life histories are available)
Poisson model is an approximation to the Markov model
where Ex is considered to be ﬁxed over the year
When suﬃcient data are available (ie full information about
the total waiting time), the two-state Markov model is
preferred because using the binomial model will not make the
fullest use of the information. 26/30 Actuarial Statistics – Module 6: Parametric models: Binomial and Poisson models
Discussion: Binomial, Poisson and Markov multi-state models Estimating parameters
if exact dates of birth, entry, exit and death are known, then
the MLE of the Markov model is easily calculated
the binomial model is complicated and need further simplifying
assumptions (eg, Balducci assumption) when µ is very small, the actuarial estimates provides
when µ is small, the diﬀerences between the two-state Markov
models and Poisson models are tiny
In practice, there i...
View Full Document
This document was uploaded on 04/03/2014.
- Three '14