A tale of two pension plans: Measuring pension plan risk from an economic capital perspective

Years of high inflation, good investment returns and profits during the 1970s and 1980s created the illusion that defined benefit (DB) pension plans are easily affordable. Due to the creation of large surpluses during those years, pension risks have generally been excluded from an organisation's general risk management processes. Over the past decade or more however, increasing life expectancy and a steady fall in interest rates have meant that pension costs have increased. Consequently, many DB pension plans now have insufficient assets to cover all of their promised benefits. As a result, security of members' benefits may be compromised.

This research, funded by the Society of Actuaries, builds on the works of Porteous et al. (2012), who performed a risk assessment of UK's Universities Superannuation Scheme (USS) based on the 2008 USS valuation report. In this research project, we update the analysis based on the most recently available valuation report. We then extend our analysis to carry out risk assessment of a stylised US plan, with the same membership profile as USS but with plan provisions modified to reflect a typical US DB plan design.

We employ an economic capital approach to assess risks. Although the term economic capital has been widely used within the banking and insurance sectors, the concept is relatively new in the context of risk assessment of DB pension plans. In this research, we adapt the commonly used definitions of economic capital to appropriately capture the specific risk characteristics of DB pension plans.

The analysis was carried out using stochastic economic scenario generators (ESG) calibrated to the UK and US economies. Specifically, we use a graphical model approach to ESG, proposed by Oberoi et al. (2019), alongside the well-known Wilkie model, to capture the sensitivity of the results to the choice of ESGs employed. The analysis also used a stochastic mortality model, similarly calibrated to the UK and the US.

Results are shown for the full distribution of outcomes, but emphasis is given for certain percentile levels in line with the selected degree of confidence.

We find that as a percentage of starting assets, the US stylised plan is more volatile than the USS plan. Moreover, the benefits of a larger allocation to long bonds are greater in the US stylised plan than the USS plan. In general, the effect on economic capital (for both plans) is much larger for changes in asset allocation than for changes to plan contributions.

The full distribution of results provided should assist plan sponsors to understand the full range of uncertainties that they are assuming in the financing of their DB pension plans. An economic capital framework provides pensions regulators with another tool to consider their exposure to benefits guaranteed by the Pension Protection Fund and the Pension Benefit Guaranty Corporation. The analysis can also help the DB pension plan members to understand the uncertainties that the sponsor faces in the financing of DB pension plans.