Title: ALM under distributional uncertainty: From DSP to DRO optimal pension fund management
Authors: Asmerilda Hitaj - University of Insubria (Italy) [presenting]
Abstract: The focus is on a second pillar defined benefit (DB) occupational pension fund (PF) asset-liability management (ALM) problem from the perspective of a PF manager delegated to pay benefits to the employees, the PF members by a company, and the sponsor who is also funding the pension plans. The pension fund collects the contributions from the sponsor and pays the benefits to the passive members. We do not consider the possibility of covering the fund through an insurance company (which certain systems, is compulsory). The PF manager's objective is to determine an investment strategy that allows the fund to cover its liabilities while minimizing the cost of funding, given by the contributions paid by the sponsor and the deficit between liabilities and total asset value at the end of the time horizon. We take into account uncertainty over members' lifetime and asset returns. In particular, the LeeCarter model is considered for the survival probabilities and the Nelson/Siegel one for the yield curve. The ALM problem is formulated in constant monetary values to immunize the impact of inflation. To solve the problem, we propose a distributionally robust stochastic optimization (DRSO) approach and analyze how the choice of metric affects the worst-case distribution and the out-of-sample performance of the solution.