Volatility in the endgame – friend or foe?

The idea

The essential idea is that, in many cases, mismatching buyout sensitivity provides trustees with the option of being able to capture buyout sooner. If buyout pricing improves because credit spreads widen (buyout prices tend to improve when credit spreads widen because insurers invest in credit to support buyout contracts), this can provide an opportunity to lock in favourable buyout pricing. However, the situation can be asymmetric because, should the opposite occur, the scheme may not have to do anything. The scheme might be able to:

• Wait for credit spreads to mean revert (which likely requires a few years) and/or
• Wait for the duration of the liabilities to shrink, so the level of the buyout spread doesn’t matter as much. In the limit, the scheme is just ‘run off’ in a low dependency or self-sufficiency strategy.

The second of these is more reliable, but requires the patience and ability to run the scheme off if need be. This may not actually be possible if the scheme has a weak covenant, for example.

The question is: how valuable is this mismatch, and could it be significant enough to warrant eschewing credit-heavy strategies in the endgame?

Strategies considered

To make a fair comparison we considered two different investment strategies that are each expected to return gilts + 0.5% per annum.

(A) A 50% allocation to credit (assumed excess returns over gilts of 1% p.a.) with the remaining 50% in cash and liability-driven investments (LDI). This is assumed to be a ‘buyout-aware’ strategy in the sense that moves in asset prices are well aligned with changes in buyout liabilities. It is also broadly like a cashflow-driven investment (CDI) strategy, designed for self-sufficiency.

(B) A 12.5% allocation to diversified growth (with assumed excess returns of 4% p.a.) with the remaining 87.5% in cash and LDI. This strategy makes the most of multi-asset diversification.

The cash and LDI are used to maintain full hedging of interest rate and inflation risk.[1]

Trigger happy

We considered a simple setup where the liabilities consist of a fixed bullet cashflow due in 20 years. The scheme is initially c.89% funded on a solvency basis (i.e. a buyout basis).[2]

To start, we didn’t model any buyout trigger, so if the scheme reaches 100% funding it doesn’t do anything even though it could secure all benefits.[3] The figure below shows the funnels of doubt on a solvency basis under strategies (A) and (B). We’ve plotted the 10^th and 90^th percentiles of both strategies, corresponding to the 1-in-10 best and worst outcomes on a solvency basis, together with the median.

Over the period until the liability cashflow is paid (20 years), strategy (A) beats strategy (B) as it has the same expected return but a narrower range of outcomes. The assumptions we’ve made are in line with the idea of cashflow matching and CDI being an efficient way to meet cashflows for relatively low return targets.

But what if we allow for a path-dependent trigger? The next picture shows what happens if we assume that the scheme buys out as soon[4] as it reaches 100% funding on a buyout basis.

As you can see, it makes a big difference!

The benefit of a trigger is that it eliminates those scenarios where you reach full funding but don’t act and the funding level then falls. This improves outcomes for members more under strategy (B) than under strategy (A) because the former is more volatile relative to buyout prices. The 1-in-10 downside outcomes (the lower lines) are improved such that, by 20 years, strategy (B) beats strategy (A).

We also see a much higher chance of potentially being able to buy out in the next five years, jumping from negligible under strategy (A) to 40% under strategy (B). The table below summarises some of the statistics.

No free lunch

There is no free lunch, of course. In the short term, uncertainty in the buyout funding position is much greater. This is a potential issue should the scheme be forced to buy out in the near term, for example due to sponsor insolvency. It is also the case that deeper in the tail (the 1-in-20 downside), outcomes are worse under strategy (B) even if the scheme is fully run off.

Mismatching merits

This all said, our analysis suggests that a strategy of deliberately not investing like an insurer can have significant value. This isn’t to say the way insurers invest is bad – far from it. Rather, the mismatch provides a potential option that could be surprisingly valuable for some schemes, particularly in the following circumstances:

• The chance of forced buyout due to sponsor insolvency is low in the short term.
• The duration of the liabilities is relatively high so that the volatility in the buyout funding position caused by moves in credit spread is higher.
• The scheme is relatively well funded so that there is a reasonable chance of the buyout trigger being met.
• At very high funding levels, the buyout trigger is monitored (and executed) on a frequent basis so that full funding can be caught quickly.

 Practical implications

Watch this space for more research in this area, including:

• How to implement buyout price monitoring to act as quickly as possible once full buyout funding is achieved.
• The potential benefits of a phased approach to capture buy-in pricing over time, rather than only insuring all at once.
• How to deal with the fact that different insurers have different buyout price sensitivity.

[1] Allowing for any implicit hedging from return-seeking assets.

[2] Please let us know if you would like details on our other model assumptions.

[3] This is a common assumption in models that can’t deal with path dependency.

[4] Within a month. Please see this blog for a discussion of the importance of frequent monitoring.