Abstract

We establish 16 good practice principles for modelling defined contribution pension plans. These principles cover the following issues: model specification and calibration; modelling quantifiable uncertainty; modelling member choices; modelling member characteristics, such as occupation and gender; modelling plan charges; modelling longevity risk; modelling the post-retirement period; integrating the pre- and post-retirement periods; modelling additional sources of income, such as the state pension and equity release; modelling extraneous factors, such as unemployment risk, activity rates, taxes and welfare entitlements; scenario analysis and stress testing; periodic updating of the model and changing assumptions; and overall fitness for purpose.

Highlights

  • If a defined contribution (DC) pension plan is well designed, it will be a single, integrated financial product that delivers, at reasonable cost to the plan member, a pension that provides a high degree of retirement income security

  • We have spent over two decades thinking about the design of DC plans as well as the modelling of different aspects of the design and, over the course of this work, we have conceived and built a DC pension simulation model called PensionMetrics

  • A good model should consider the full set of choices that are available to the plan member

Read more

Summary

Introduction

If a defined contribution (DC) pension plan is well designed, it will be a single, integrated financial product that delivers, at reasonable cost to the plan member, a pension that provides a high degree of retirement income security. Model Specification and Calibration Any DC pension model should be built using plausible assumptions about the stochastic processes driving key variables (such as asset prices, interest rates and mortality rates) and these processes should have empirically plausible calibrations These calibrations would cover, inter alia, the risk premia on growth assets such as equities, the interest-rate process and the mortality process, across the whole time horizon relevant for the plan member. The UK regulator, the Financial Conduct Authority (FCA), requires plan sponsors to provide deterministic projections of pension fund values for hypothetical pension fund returns of 2%, 5% and 8% (Financial Conduct Authority 2021). Cox-Ingersoll-Ross model (Cox et al 1985), and mortality rates are modelled using the age-period-cohort factor to cover adminsitsotcrhaatsitoicnmcohrtaalritgyemsoadneldwphircohvoridigeinratpedroinfiHt.oIbncrvaeftsettmal.e(n19t8r2e).turns are modelled using a multivariate Gaussian process and calibrations set out in Harrison et al (2012), interest and inflation rates are modelled using the Cox-Ingersoll-Ross model (Cox et al (1985)), and mortality rates are modelled using the age-period-cohort stochastic mortality model which originated in Hobcraft et al (1982)

Modelling Member Choices
Modelling Member Characteristics
Modelling Plan Charges
Modelling Longevity Risk
Modelling the Post-Retirement Period
11. Modelling Extraneous Factors
13. Periodic Updating of the Model and Changing Assumptions
14. Fitness for Purpose
15. Attractiveness of the Approach
16. Conclusions and Caveat
Principle 3
Principle 7
10. Principle 10
14. Principle 14
16. Principle 16
Findings
Optimise attention
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call