The Implementation of the Multiperiod Mean-Variance for Asset Allocation of Pension Fund

Abstract

This paper investigates the application of multi-period mean-variance methods for optimal asset allocation in pension fund institutions by considering mortality factors and changes in contributions. The selection of assets used to find portfolios is using Treasury Bills as a risk-free asset and using stock included in the LQ45 index as risk assets that have the best performance during the period 2014 -2018. The multi-period mean-variance method's main purpose is to obtain the minimum variance with optimal returns at the end of the period. This result indicates that the smallest return is owned by the asset allocation model without contribution because, without additional membership fees, the amount of funds that can be invested is more limited. Hence, the ex-pected wealth terminal is smaller than other asset allocation models. The asset allocation without the mortality factor gives higher expected terminal wealth compared to other asset allocation models. This is because the absence of mortality factor in the pension fund investment will not be forced to stop before the retirement pe-riod ends so that the portfolio return obtained is higher than other asset allocation models. While the asset al-location model using the risk-free asset as references has a higher risk with the same return compared with as-set allocation using risk assets as a reference. This indicates that the asset reference used affects the return and risk on the allocation of assets of the Pension Fund at the end of the period investment.