An intelligent asset rebalancing strategy can help to reduce pension scheme risk. George Tyrakis discusses the possible approaches and potential pitfalls

Traditionally pension schemes have maintained static asset allocations, where the portfolio is allocated across a range of asset classes in line with the scheme's risk budget. As asset values fluctuate, the asset portfolio requires rebalancing to ensure the desired portfolio weights are maintained. Naturally, a number of questions arise. For example, at what frequency should the portfolio be rebalanced? What are acceptable limits for the portfolio weights to deviate from the initial allocations? Is a fixed asset allocation suitable for the scheme at future dates?
Using an Asset-Liability Model with a suitably calibrated Stochastic Economic Scenario Generator (ESG), the assets and liabilities of the scheme have been projected stochastically over future periods. The yield curves used to discount the liabilities were produced in the ESG and these were also used to drive returns on fixed income assets. Real-world correlations between various asset classes were modelled together with the realistic dynamics of the assets held, such as equity market jumps and time-varying equity volatility. For the purposes of this exercise, longevity risk was not modelled so as to focus the analysis on market risks and how these impact the assets and liabilities. The asset portfolio used is shown in Table 1.

Possible outcomes
The range of possible funding level outcomes over the time horizon can be illustrated using a probability distribution funnel generated using Monte Carlo simulation.
In Figure 2, assets are rebalanced on a monthly basis using the initial allocation as a target benchmark. This can be thought of as 'strategy neutral' because the asset allocation is mechanically rebalanced on a passive basis.
Consider revising the asset portfolio rebalancing strategy, so that the target asset allocation changes as the funding level of the scheme moves into different ranges as shown in Figure 3.
If the funding level breaches any of the conditions shown above, from above or from below, the asset portfolio is rebalanced accordingly. Under this strategy, the initial asset allocation (shown in the Funding Level <85% column of Figure 3) requires a small increase in UK equities to in order to off set some of the lost return due to de-risking so as to ensure that the scheme's target of achieving return due to de-risking so as to ensure that the scheme's target of achieving 100% funding in three years is attained on average.

? Probability of Success, defined as the probability of being at least fully funded over the three years, is relatively unchanged under both strategies given both strategies target and achieve a median funding level of 100% at the end of three years.
? Standard Deviation of funding level is reduced by 5% per annum.
? Conditional Expectation, defined as the expected funding level given the funding level is below 100% in year three, has increased by 2%, another desirable outcome.
? The potential deterioration of funding level over time is significantly reduced as can be seen by lower 1-in-200 Value at Risk and Expected Shortfall risk measures.
The above reduction in risk has come at the expense of reduced upside potential. The gradual risk reduction rebalancing strategy is effectively reducing the overall riskiness of the asset-liability position, and therefore its return potential. Despite this, a more closely matched asset-liability position over time would be desirable from a funding perspective, as it would increase the probability of achieving the investment objective, which has previously been illustrated (Puchy, 2012).



BIBLIOGRAPHY
Puchy, R. (2012, October). Barrie & Hibbert.
R. C. Urwin, S. J. (2001). Risk Budgeting in PensionInvestment. British Actuarial Journal, 319-364.