Any reforms of pension systems inevitably involve their optimization, which is a challenging task since pension systems are dynamic, multidimensional and are affected by a variety of demographic, investment-related, and institutional random impact factors. The model described in this article aims at demonstrating the dependence of the target functions of pension systems on such factors. The current research sheds light on the influence of demographic parameters on funded and unfunded pension systems and shows the importance of institutional risks in both types of systems. The values of the state-regulated parameters for 2030 are specified, which allows us to maximize the key target functions: the replacement rate and pension benefits. Further, the results of empirical analysis of the impact factors affecting pension systems of OECD countries are described. The novelty of the paper lies in the analytical and quantitative methods used for the optimization of the pension system on the basis of the replacement rate.