Many defined benefit schemes are in a position where they experience negative cashflows. This happens because each month or year as they pay more in benefits than they receive in contributions. There is nothing wrong with this, of course. In fact, it is expected as pension schemes mature and fulfill their purpose of using the assets to pay members’ benefits.
Other investors can also be in a similar position – DC schemes in decumulation, endowments and even sovereign wealth funds.
Investors in this position can face a number of challenges compared to schemes in otherwise similar situations, but with a lower net cashflow out. These challenges can fundamentally change the nature of these funds and makes them a very different animal to funds with zero or positive cash flow. This necessitates quite a different approach to risk management and asset allocation.
We call this “End Game ALM”.
The Pensions Regulator specifically highlighted this issue and the need to plan for it in its 2016 Annual Funding Statement.
To help trustees understand and assess the potential impact, we can quantify some of these challenges using ALM tools & measures. By doing this, ultimately, trustees will be able to make better decisions that take the unique characteristics of the scheme into account.
We liken this to the Windchill Factor. It’s like being at the top of a hill: while the ambient temperature might be a moderate 10 degrees or so, the effect of windchill means that the actual temperature you experience might be close to zero. If you’re planning a long walk in those conditions, you better factor that windchill into your planning.
It’s the same with End Game ALM.
While a standard risk measure (volatility say) might tell you the risk in the portfolio is a moderate 10%, as the effect of negative cashflow is taken into account it could be that this increases by as much as 2x. This is caused by the sequencing risk associated with a negative cashflow profile – the sequence of returns matters when you are taking money out.
For example, it isn’t necessarily enough for assets to earn RPI+3% over 20 years if the assets take a heavy setback early on, and assets are also then liquidated to pay benefits. This can put the scheme into a position from which it is hard to recover. This effect isn’t captured by standard risk measures such as VaR or volatility.
A Tale of 2 Schemes
Let’s illustrate this with an example comparing two different theoretical pension schemes with identical funding levels and asset allocation:
1. New Industry Inc. Pension Scheme
The bulk of this scheme’s liabilities relate to deferred members. It closed to future accrual relatively recently, and many of its members are in their 30s and 40s.
2. Old Industry Plcs. Pension Scheme
This scheme has most of its liabilities in relation to pensioners and older deferred members. It closed to future accruals many years ago.
There are two main impacts of the shorter, more drawdown-heavy cashflow profile of Old IndustryPlc:
- It generally needs a higher Required return than a less mature profile, given a fast dwindling asset pool with assets sold to pay benefits (other things being equal), which means the remaining assets have to work harder
- It generally leads to a higher chance that the Required return in the future may increase substantially (we call this Required-Return-At-Risk “RR@R”), due to the sequencing effect of suffering an asset loss, followed by the need to pay a significant cashflow. This effect means, even with an equivalently-risky asset portfolio, a more mature scheme is more likely to suffer significant setbacks, which may make the required return unachievable.
These effects are illustrated below:
|Case 1||Case 2|
|Required Return to 2040 Gilts +||Gilts +2.2% p.a.||Gilts +3.2% p.a.|
|VaR 95% (£m)||£1.57bn||£1.54bn|
|Required Return following adverse event (95% 1-year)||Gilts +3.3% [+1.1%]||Gilts +5.0% p.a. [+1.8%]|
The table above shows that the two schemes may superficially appear to be in quite similar positions, when conventional risk metrics like VaR and FRaR are analysed, these give very similar results.
However, when we look at risk measures that take the cashflow profile into account, this materially increases the riskiness of the position faced in the second case.
What the final row in the table tells us is that the pension scheme in the second case is at a heightened risk of ending up in an unsustainable situation; while returns of gilts +3%p.a. might well be viewed as achievable over a long time horizon, returns in excess of gilts +5%p.a. are much tougher to achieve while keeping risk within realistic bounds.
So what do you do now?
There are some important consequences of these results.
First, it’s important to know if your scheme is in a negative cashflow situation at the moment.
If it is, you may need a different set of risk lenses in order to properly evaluate the risks in the scheme. VaR and volatility are less useful in this position, instead, a focus on Required Return and Required Return at Risk can help.
If it isn’t in the position of significant net cash outflow right now, almost all pension schemes will be at some point. Of course, there is nothing wrong with this, but it is important to plan with this point in mind.
One metric that can help is to project the cash outflow as a percentage of the projected future asset value. You can investigate when this exceeds a certain level (call this the “Relative Cashflow Threshold” or RCT). The scheme may well take the view that it would like to be fully funded before the RCT exceeds a certain level, and plan its investment strategy, and sponsor contributions accordingly.