Measuring and explaining implicit risk sharing in defined benefit pension funds

Abstract

This article investigates responses to changes in solvency by occupational pension funds using a unique panel data set containing the balance sheets of all registered pension funds in the Netherlands over a period of 13 years (1993–2005). A fixed discount rate for liabilities in the supervisory framework allows us to measure the response of pension funds to solvency shocks. We find that pension rights are expanded, by e.g. indexation, or limited, by for instance setting the pension premium over its actuarially fair price, in line with the funding ratio but that the pension funds’ response function exhibits two sharp and significant behavioural breaks, close to the minimum funding ratio of 105% and the target ratio of around 125%. We further find that large funds and grey funds are relatively generous to current participants.

Keywords:

• defined benefit pension funds
• risk sharing
• funding ratio
• indexation
• recovering behaviour

JEL Classification:

• G23
• G28

Acknowledgements

The authors are grateful to Dirk Broeders, two anonymous referees and participants of the 15 November 2010 Netspar conference in Maastricht and the 27 January 2011 DNB research seminar. The views expressed in this article are personal and do not necessarily reflect those of CPB or DNB.

Notes

¹ Until 2004, pension funds had to report to the ‘Pensioen- en Verzekeringskamer’. This supervisor merged with the Dutch central bank in 2004.

² A more detailed description of how we handle mergers and cessations is available from the authors upon request.

³ Backservice is the increase of the existing pension rights due to a change in the actual wage of the participants in a final wage scheme.

⁴ Investment portfolios with relatively safe AAA-bonds are safer than equity portfolios, but for our financial soundness indicator we consider only the funding ratios. Investment risk has not been considered in our analyses.

⁵ The current funding ratio is also problematic from an econometric point of view as this choice would result in a spurious regression. Indeed, unexpected changes in the level of fund obligations (through e.g. demographic realizations or decisions about indexation) would show up in both the funding ratio and the transfer rate.

⁶ The required funding ratio of around 127% is determined so that the probability of underfunding within 1 year is limited to 2.5%.

⁷ This jump can be significant. Suppose that for a pension fund 40% of the technical reserves are due to the pension rights of active participants, whereas the rest are provisions for the currently retired and deferred. Assume that individual wages grow with 5%. This implies that the technical reserves of this fund under a final wage scheme increase by 2%, just because of the backservice. Switching to an average wage scheme would eliminate both the backservice and the increase and would result in a that is 2 percentage points lower. Once the fund has implemented an average wage scheme, it has much more control over its technical reserves as all indexation is, within limits, at the discretion of the board.