A Duration Dependent Multistate Lifetable with Applications to Nuptiality

Juha Alho, University of Joensuu

Multistate lifetables provide rich opportunities for describing complex population systems. However, their practical application has been limited by data problems, and by strong Markov assumptions. We will describe a program MTABLE that is capable of implementing a multistate population system based on age and duration specific estimates of transition intensities between the states. The program uses the so-called Runge-Kutta method to calculate the probabilities of transition from state to state. In the duration dependent case the lifetable calculations are carried out by stochastic simulation. MTABLE can be used to calculate the distribution of the population by the time spent in a specified state. This permits the calculation of other moments, in addition to the expectation. MTABLE can also be used to calculate probabilities of exiting a particular state via another one. The use of the program will be illustrated with nuptiality data for the Finnish women in 1998.