Media On the Optimal Design for a Collective Defined Contribution Fund: A Bayesian Optimization Approach

On the Optimal Design for a Collective Defined Contribution Fund: A Bayesian Optimization Approach

uploaded April 16, 2021 Views: 81 Comments: 1 Favorite: 0 CPD
Speakers: 
Description:

Typical pension funds with multiple cohorts disconnect the generational contribution and the generational benefit. On the one hand, the individual right is unclear in such a collective design, and on the other hand, the fund might encounter the endogenous mismatch of cash flows due to the underlying environmental change (e.g, demographic change). In this paper, we provide a novel design for the defined contribution (DC) pension fund with multiple cohorts. The assets from all generations are managed together while the individual benefit account for each generation is established via the redistribution of the pension asset. Through the return-smoothing mechanism, intergenerational risk sharing (IRS) is explicitly implemented in a DC fund. We use Bayesian optimization to search for the optimal asset-liability management rule for the fund numerically. Our simulation study shows that the development of a benefit account in our fund is much smoother than a pure defined contribution (DC) plan where IRS is not implemented. Further, our fund improves the welfare of participants in tough markets (with low market price of risk) compared to a pure DC plan.
Authors: An Chen, Motonobu Kanagawa, Fangyuan Zhang
 

Tags:
Categories: PENSIONS
Content groups:  content2021

1 Comments

gg
gg

1054 Days ago

A very simple model but it has everything it should have. I liked especially the Bayesian optimisation. It is a tool which should be known and used much wider in the community.