Con Keating looks at portfolio construction for members nearing retirement and he doesn’t like what he finds.
In the course of recent work on value-for-money metrics, I came across several schemes which apply ‘life-styling’ to older members. Many have argued that these strategies merit a benchmark which differs from the traditional 80/20 equity debt construction, which reflects the overall fund market capitalisation or available investment opportunity set.
The heart of a value-for-money metric is a cost benefit or input-output comparison. In the case of defined contribution (DC) funds, the inputs are the contributions made and their timings, while the output is the market value of the fund at measurement date. This delivers the actual performance as an internal rate of return (IRR). An IRR is a risk-experienced average return. It has experienced the good and bad times of market behaviour over the period of saving. This renders otiose the confusing discussions of risk management, and the role of risk in the risk-return equation.
This requires a little explanation. One of the few things we know about risk is that it means more (bad) things may happen than will. The essence of good active fund management is the identification and avoidance or mitigation of those bad events which will occur from among those which may. Passive investment is acceptance or tolerance of whatever comes to pass.
The IRR of a member’s savings reflects the strategies followed by the funds in which it was invested. This might be the passive ‘non-strategy’, or market-timing, or lifestyling. Each have their own ex-ante risk exposures, but the IRR explicitly captures all those risks which eventuated. It also captures the costs associated with risk avoidance and mitigation of those events which did not occur.
The arguments that a value-for-money metric should reflect the ex-ante risk-preferences of the investor/fund manager are a canard. The implementation of strategies reflecting these preferences may, and usually do, restrict the value for money achievable by a saver.
The expression ‘life-styling’ refers to portfolios which progressively move their asset allocation away from equity to debt as the member approaches retirement. The objective is to minimise the volatility of the portfolio’s value at retirement, which of course, was previously the prime determinant of the member’s retirement income.
There are difficulties with creating benchmarks for ‘life-styling’ in that this practice applies at the level of the member’s allocation, not the overall consolidated fund. It is also far from clear why this strategy should be treated differently from any other strategy where comparison is made to the available investment opportunity set.
The process of moving progressively to higher bond allocations will carry with it the prospect of a smaller ‘pot’ value at retirement than would be expected with the standard 80/20 portfolio. The question then becomes, is this lower pot value warranted by the risk avoided?
This may be investigated quantitively. I set up a simple model, where debt securities have an expected return of 4% and volatility of 10%, with equities having an expected return of 7% and volatility of 20%. The correlation between them is 0.4. The returns distributions are assumed to be normal. The portfolio allocation is then modified from an initial 80/20 to all bonds in steps of 5% until it reaches 100% debt after 17 years.
At this time the expected value of the 80/20 benchmark is some 20.45% higher than the ‘life-styled’ fund. The likelihood of the 80/20 benchmark delivering a return below the ‘life-style’ expected value is 14.24% and the expected loss relative to that value is 7.63%. By contrast, the life-style fund has a 50% likelihood of a return below its expectation with a value of 6.74%. If we define risk as the product of the likelihood and magnitude of an event, then the risk of the ‘life-style’ strategy is 3.37% while that of the benchmark 80/20 portfolio is just 1.08%. It appears that in this instance ‘lifestyling’ is a case of reckless prudence in risk management.
The loss events of the ‘life-styling’ and benchmark portfolios are not independent of one another; they have a degree of dependence by construction. 32.8% of the benchmark portfolio’s problematic returns will be associated with underperformance by the ‘life-style’ portfolio; a case of damned if you did or damned if you did not follow the strategy.
In the vast majority of outcomes, the pension saver is far better off invested in the 80/20 benchmark; the ‘lifestyle’ portfolio only exceeds the expected value of the benchmark 80/20 portfolio in 2.04% of circumstances.
Of course, this is one simple illustration and we might vary any or all of the model assumptions, the expected returns and volatility of debt or equity, their correlation or the speed at which ‘life-styling’ is introduced. We might even introduce more complex rules, path-dependent strategies, such as moving to bonds only after a strong equity return, but these will all bring with them only variations in degree of the problem illustrated here.
If we add to this concerns that we may currently be in a debt bubble induced by monetary and quantitative easing, with the implications of that for future debt returns and volatility, ‘life-styling’ appears to be far from conservative and, indeed, unlikely to deliver the benefits usually claimed for it
