Tvaraus pensijos portfelio sudarymas dinaminiu stochastiniu imitaciniu modeliavimu

Abstract

Due to the instability of world economy, increased longevity, higher healthcare and long-term care costs, an adequate retirement income level is a daunting challenge today. A sharp drop in equity markets in 2008 has reduced both private savings and pension fund assets. Today’s retirees have to make increasingly complex financial decisions. Gone are the days when one could rely solely on the social security system. The increasing topicality of sophisticated retirement planning is obvious, as retirees will soon be expected to fund larger portions of their retirement spending. The questions of retirement portfolio management will gain focal attention. While various measures can be undertaken to accommodate the individual needs, goals and wishes of today’s retirees, the article focuses on how smart portfolio construction can extend the longevity of retirement savings and even enhance the returns.The article reviews the results of former pieces of research in the field of retirement portfolio formation and discusses several methods to determine asset allocation for a retirement portfolio, including the heuristic and the multiple horizon approach and stochastic optimization methods. In this paper, dynamic stochastic simulation and stochastic optimization techniques in Monte Carlo simulation models created by the author are applied to identify the optimal portfolio allocations for minimizing the probabilities of depleting the retirement portfolio earlier than the planned retirement horizon (often termed portfolio ruin or shortfall risk) by making constant inflation-adjusted withdrawals. The analysis of the methods to determine asset allocation for a retirement portfolio showed that an optimal portion of stocks with 5.5 per cent and higher withdrawal rate is alwayshigher than that determined using the heuristic or the multiple horizon approach. Stochastic optimization in all cases provided the least risky portfolios in terms of shortfall risk. In case the stochastic optimization is unavailable, the research results suggest applying the multiple horizon apporach for portfolios with a longer that 20-years investment horizon and a higher than 4.5 per cent withdrawal rating. Otherwise, the heuristic method should be applied.In this research, it was supposed that the retirement planning horizon was fixed. In reality it is impossible to predict the exact lifespan of a person; therefore, the stochastic models should incorporate the consideration of stochastic lifetimes as well.