Joined-Up Pensions Policy in the UK: An Asset-Liability Model for Simultaneously Determining the Asset Allocation and Contribution Rate

Abstract

ABSTRACT:The trustees of funded defined benefit pension schemes must make two vital and interrelated decisions -setting the asset allocation and the contribution rate. While these decisions are usually taken separately, it is argued that they are intimately related and should be taken jointly. The objective of funded pension schemes is taken to be the minimization of both the mean and the variance of the contribution rate, where the asset allocation decision is designed to achieve this objective. This is done by splitting the problem into two main steps. First, the Markowitz mean-variance model is generalised to include three types of pension scheme liabilities (actives, deferreds and pensioners) In a funded defined benefit scheme the employer and employees both make contributions to a fund which is invested to provide the pension, and any other benefits due under the scheme. The benefits received under such schemes are defined in advance, usually as a proportion of the employee's final salary. Many UK companies have recently chosen to close their defined benefit pension schemes. In the 5½ years up to February 2003, 63% of UK final salary schemes were closed to new entrants, while an additional 9% of schemes were also closed to future accruals (Association of Consulting Actuaries, 2003). The reasons given for closure include the introduction of Financial Reporting Standard 17, the substantial deficits on final salary schemes (caused by the fall in interest rates, the major stock market decline after the peak in December 1999, the extended contribution holidays and contribution reductions for employers, increases in benefits, the conversion of discretionary benefits into non-discretionary benefits, the use of pension schemes to finance early retirement on very favourable terms, and the tax limit on scheme surpluses); the effective move from limited price indexation to fully indexed pensions, with the fall in annual increases in RPI to below 5% since July 1991; the regulatory burden of administering these schemes; the increased cost due to rising life expectancy; the increased size of pension liabilities, relative to the size of the employer; the increase in stock market volatility; the risks that such schemes impose on employers (e.g. the risk that the fund will be insufficient to pay the pensions, the credit rating of the employer may be reduced because of the possibility of pension shortfalls); the abolition of tax relief on dividends from UK companies in 1997; the changes in actuarial technique leading to more volatile surpluses; the risk that new legislation or decisions by the law courts will increase the liabilities; the lower priority given to retaining staff; the opportunity to establish defined contribution schemes with a lower cost to the employer, and the much greater portability of defined contribution schemes.This paper develops an approach to the simultaneous analysis of two critical and interrelated decisions which must be made by any fund's trustees: the fund's asset allocation and its contributio...

Asset–liability modelling and pension schemes: the application of robust optimization to USS

“…These monthly ratios were averaged to give the mean funding ratio. The mean projected contribution rate was computed using the actuarial formulae in Board and Sutcliffe (2007) with a spread period (M) of 15 years (see Appendix A). The number of years accrued by the average member (P) was 18 years, while administrative expenses were set to zero.…”

“…We compare these results with those of four benchmarks. The first is the actual portfolios chosen by the Universities Superannuation Scheme (USS), and the second is a modified version of the Sharpe and Tint (1990) model which has been widely used in ALM models (Ang, Chen & Sundaresan, 2013;Board & Sutcliffe, 2007;Chun, Ciochetti & Shilling, 2000;Craft, 2001Craft, , 2005De Groot & Swinkels, 2008;Ezra, 1991;Keel & Muller, 1995;Nijman & Swinkels, 2008). The third is Bayes-Stein estimation of the inputs to the modified Sharpe and Tint ALM model, and the last is Black-Litterman estimation of the portfolio inputs to the modified Sharpe and Tint ALM model.…”

“…Since the assumptions for a theoretical justification are more or less satisfied and the multi-period alternative is computationally impractical, a single period model appears to be a reasonable simplification and has been used by many researchers for solving pension ALM problems; including Ang, Chen & Sundaresan (2013); Board and Sutcliffe (2007); Chun, Ciochetti & Shilling (2000); Craft (2001Craft ( , 2005 Ultimately pension scheme trustees are concerned about the scheme's funding ratio and contribution rate. Since the asset allocation has an important influence on the contribution rate and funding ratio, trustees need to make the asset allocation decision in the light of its effect on these two variables.…”

“…MacBeth, Emanuel and Heatter (1994) report that pension scheme trustees prefer to make judgements in terms of the scheme's funding ratio and contribution rate, rather than the scheme's asset-liability portfolio. We use the model developed by Board and Sutcliffe (2007), which is a generalization of Haberman (1992), to transform the asset allocations to projected contribution rates. This generalization allows the discount rate to differ from the investment return, which improves the economic realism of the actuarial model.…”

The Cult of the Equity for Pension Funds: Should it Get the Boot?

“…Boender (1997) used data for a Dutch pension scheme and found that a 10% equity allocation minimised risk. Mulvey and Thorlacius (1998) and Board and Sutcliffe (2004) used hypothetical data and found the riskminimising portfolio had 18.7% equities and 8% equities respectively. Another three studies used data on a Dutch pension scheme to find an optimal asset allocation.…”