Optimal early retirement near the expiration of a pension plan (Q854273)

The author develops a model for the financial valuation of the retirement benefits. He considers a pension plan with the option of early retirement, where the benefits, based on salary, are paid at the time of retirement. Let \(S_t\) be the salary of a pension plan member at time \(t\), then the retirement benefits are \(\psi(t,S_t)\) in the case of retirement of the member at time \(t\). Assuming that \(S\) is a Markov process, solution of a stochastic differential equation, a variational inequality for the financial value \(V\) of the retirement benefits was previously obtained, the existence and uniqueness of the solution was established, and its free boundary \(\Gamma\) was studied. In this paper, the behaviour of \(\Gamma\) is discussed near the expiration of the pension plan. The main idea is to compare \(\Gamma\) with the free boundary associated to a specific option. It is established that if the limit of the free boundary is a point of regularity of \(\psi\), then the convergence rate is parabolic, otherwise a logarithmic factor appears.