Two approximations of the present value distribution of a disability annuity (Q2571228)

Deriving the distribution of a complex life or disability insurance contract can be quite complicate. For instance, under the Markov assumption, it involves solving a set of integral equations. It is often impossible to obtain a closed form solution for the conditional probabilities. One approach is to replace the variable of interest by an another one that is stochastically dominated but whose structure is simpler. This new variable is much more tractable analytically, it is riskier therefore providing for an upper bound to the effect of uncertainty, and its distribution may be actually quite a close approximation to the one of the original variable. This paper applies this approach to the framework of three state disability insurance. It constructs the simplifying variable by assuming that the disable state is absorbing. When the time until death is used as an additional conditioning variable, simulations shows that the upper bound derived is a very good approximation of the original variable.