Intergenerational risk sharing in a defined contribution pension system: analysis with Bayesian optimization
This paper considers a fully funded, collective defined contribution (DC) pension system in the presence of multiple overlapping generations. The focus is to investigate whether the welfare of participants may be improved by intergenerational risk sharing (IRS). To this end, a realistic investment strategy with no borrowing is implemented in the absence of external financing of the pension fund. Specifically, in the implementation of the IRS, the pension system adopts an automatic adjustment rule for the indexation of individual accounts adapting to the notional funding ratio of the pension system. Two parameters in the pension system determine the investment strategy and the strength of the adjustment rule, which are optimized by expected utility maximization using Bayesian optimization. The authors show that the trade-off between the volatility of the retirement and that of the funding ratio can be controlled by optimal adjustment parameter to attain the IRS. They also show that compared with the optimal individual DC benchmark using the life-cycle strategy, the pension system with the IRS can improve the welfare of risk-averse parameters when the financial market is volatile. Section 2 of the paper describes the IRS-DC pension model, where the pension fund contains multiple overlapping generations. The analysis of the IRS-DC pension model is provided in Section 3. Two key results are presented. Specifically, the first result presented in Proposition 1 describes the impacts of the adjustment parameter on the funding ratio through the first two conditional moments of the log funding ratio. The second result provided in Proposition 2 describes the impacts of the adjustment parameter on the pension benefits through analysing an expression for the retirement benefit of each generation. The effects of intergenerational risk sharing are also studied. The expected utility maximization problem for optimizing the investment strategy and the adjustment parameter is formulated and discussed in Section 4. Numerical results are provided and discussed in Section 5. Specifically, several aspects, including the optimal investment strategy and adjustment parameter, funding ratio process, the dynamics of individual accounts, distribution of retirement benefits, and welfare of participants, are considered.
- A NEURAL NETWORK BOOSTED DOUBLE OVERDISPERSED POISSON CLAIMS RESERVING MODEL
- A neural-network analyzer for mortality forecast
- Bias regularization in neural network models for general insurance pricing
- Calibrating the Lee-Carter and the Poisson Lee-Carter models via neural networks
- Convergence rates of efficient global optimization algorithms
- Empirical properties of asset returns: stylized facts and statistical issues
- Gaussian processes for machine learning.
- scientific article; zbMATH DE number 51724 (Why is no real title available?)
- Investment Policy for Defined-Contribution Pension Scheme Members Close to Retirement
- Mean-variance target-based optimisation for defined contribution pension schemes in a stochastic framework
- Measuring the roughness of random paths by increment ratios
- Optimal investment for a defined-contribution pension scheme under a regime switching model
- Optimal investment strategies and risk measures in defined contribution pension schemes.
- Optimal investment strategy for defined contribution pension schemes
- Optimal management under stochastic interest rates: the case of a protected defined contribution pension fund
- Optimum consumption and portfolio rules in a continuous-time model
- Pension funds as institutions for intertemporal risk transfer
- Pension saving schemes with return smoothing mechanism
- Point and interval forecasts of death rates using neural networks
- Stochastic lifestyling: optimal dynamic asset allocation for defined contribution pension plans
This page was built for publication: Intergenerational risk sharing in a defined contribution pension system: analysis with Bayesian optimization
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6569737)
