A scheme’s liabilities can be valued in several different ways with different discount curves and spreads used depending on the purpose of the valuation. Following the volatility experienced in Gilts last year and with expectations of continued uncertainty, we have recently had client conversations about the impact of moving from Gilts- to swaps-based discounting. Since we believe this topic could be of interest to other clients, we are sharing this update, written by one of our LDI Client Portfolio Managers, Carl Svensson, more widely.
As a reminder, LIBOR benchmark rates have historically been the main interest rate benchmarks used in swap markets, but the market was reformed in the wake of the Global Financial Crisis and the relevant interest rate benchmark for Sterling swaps is now the Sterling Overnight Index Average (SONIA) benchmark. References to swap yields below are to SONIA swap yields.
What are some of the factors a scheme should consider when deciding on the discount basis?
The scheme’s actuary will in most cases be best placed to comment on scheme-specific circumstances and any regulatory implications of changing the discounting basis in a specific circumstance. When it comes to managing LDI mandates, whilst most UK defined benefit schemes now use Gilt-discounted liabilities as the benchmark for their LDI mandates, we have historically seen some UK schemes discount their liabilities on a swaps curve, and it is still the standard discounting basis in countries such as Ireland and the Netherlands.
The discount basis that is chosen becomes increasingly important for a scheme considering its target end state. We have seen some schemes whose end game is targeted towards a low dependency approach move to discounting on a credit curve. Similarly, for a scheme targeting a buyout it is worth considering that insurers use swaps to value their liabilities under the Solvency II regulation, but typically allocate a significant portion of their investments to credit and other matching assets. The implication being that buyout pricing tends to be driven by a combination of swap rates and credit spreads. This is an important consideration when looking at ways of reducing volatility to buyout pricing.
Figure 1 below sets out some of the typical, high-level considerations for a scheme when deciding on the discount basis.
Figure 1 – The importance of the liability valuation approach
This chart is not intended to be relied upon as a research or investment advice, and is not a recommendation, offer or solicitation to buy or sell any securities or financial product or to adopt any investment strategy.
Source: BlackRock as at July 2023.
What are some of the implications of discounting on Gilts versus swaps?
The impact on valuation and hedging characteristics of moving from discounting on a Gilts basis to a swaps basis will differ from scheme to scheme, but some general observations can be made. As can be seen in Figure 2, while swap yields are below their Gilt equivalents for shorter dated tenors, Gilts presently yield more at around the 8-year mark. This will typically imply that for a scheme with a liability duration longer than this we would expect to see an increase in the value of the liabilities when moving from Gilts to swaps discounting.
As an example, for an illustrative set of nominal cashflows with a duration of c. 19 years when discounted on a Gilts basis, the present value increases with c. 10% when discounting them on a swaps basis. Interestingly, because the Gilt curve is significantly steeper than the swap curve, the increase is even larger in PV01 terms, where the increase is almost 20%. As such, to maintain the same hedge ratio on a swaps basis without further spread applied would, all else equal, require more leverage and/or capital.
Figure 2 – Gilt and swap yields
The figures shown relate to past performance. Past performance is not a reliable indicator of current or future results.
Source: BlackRock as at 6 July 2023.
Are the liabilities investable and additional considerations when using swaps?
Irrespective of the discounting basis the liability cashflows are an uninvestable benchmark since there is no tradeable asset exactly matching them. Most schemes’ liabilities will run out to at least the turn of the century whereas the longest-dated Gilt ‘only’ matures in 2073. Similarly, SONIA swaps are clearable up to a similar tenor in LCH and for clients discounting on a credit curve there is a limited amount of ultra-long credit available.
Schemes looking to hedge a swaps-discounted set of liabilities using swaps as primary hedging instrument need to consider some additional practical challenges. Such a scheme needs to hold collateral alongside the swaps. The returns of this collateral can have a significant impact on the return of the asset portfolio versus the liabilities. A fully funded scheme hedging swap-discounted liabilities would enter swaps where it receives a fixed rate and pays a floating rate (SONIA) while holding the remainder of the portfolio in cash and/or a cash fund. The yield on the cash holding can in some cases be lower than the floating rate paid on the swaps. This would mean that even a fully funded swap portfolio could return less than the liabilities and it is one of the reasons we have seen many Dutch and Irish clients move away from swaps-only portfolios to using a combination of swaps and bonds to hedge their liabilities. In contrast and slightly simplified, a fully funded scheme discounting on Gilts could construct a Gilts-only portfolio with no leverage. By using redemption payments and coupons from the Gilts held, this portfolio could be used to settle the liabilities as they come due with no collateral drag. Of course, this can be complicated for schemes employing leverage where the costs of gilt financing must be considered.
When a swap is traded, a decision needs to be made between trading bilaterally, or centrally cleared. There are a range of factors to consider, such as counterparty exposure, initial margin needs, transaction costs and liquidity when deciding between these approaches. We covered this topic in more detail in the May edition of Views form the LDI Desk, where we also noted that HM Treasury have announced plans to extend the pension scheme exemption for clearing interest rate swaps until, at least, 2025. But, while schemes investing through segregated mandates have been exempted from mandatory clearing since 2012, the exemption cannot be used for fund structures, for example QIAIFs commonly used by LDI investors. As such, clients investing through a fund are not able to claim the exemption for derivatives held in these and the funds will, depending on derivatives usage, be forced to clear some of these. Similarly, larger schemes with significant derivatives usage might be in-scope for the Uncleared Margin Rules (“UMR”), which mandates them to post initial margin for bilateral trades once the exposure with a counterparty has exceed a specific level. These factors mean that for many schemes, using a significant amount of swaps requires them to hold a large proportion of the assets in cash collateral with the accompanying drag on asset returns mentioned above.
For schemes that use any form of bilateral derivatives, counterparty risk needs to be carefully considered. Importantly and in contrast to an unlevered Gilts-only portfolio, these are always present in a swap portfolio, irrespective of if the portfolio is fully funded or not. While daily collateralisation reduces the risk of mark-to-market losses on derivatives there is also the risk that exposure could be closed out if a counterparty defaulted. This risk can be mitigated in several ways including i) thoughtful ongoing monitoring and management of counterparty exposure, ii) using a diversified panel of counterparties and iii) running a ‘clean’ book whereby there is a focus on minimising not just net derivatives exposure, but gross positioning as well. We also note that some counterparties have reduced their bilateral trading footprint in recent years and the number of active counterparties has decreased, which may impact views of durability of some bilateral-focused strategies.
Are Gilt yields more volatile than swap yields?
A question that has prompted some of the discussions about changing the discount basis is if hedging on a swaps basis could reduce the funding level volatility? The short answer, from historical analysis, is ‘no’. As can be seen in Figure 3, long-dated Gilt and swap yields tend to be highly correlated and though there can be short-term dislocation between the two they tend to move in tandem over longer periods. While Gilts moved by more than swaps in the Gilt crisis of September 2022, it can be debated as to whether a similar trend would have been seen in swaps had this been the primary hedging choice of pension schemes.
Figure 3 – 30-year Gilt and Swap yields
The figures shown relate to past performance. Past performance is not a reliable indicator of current or future results.
Source: BlackRock, Bloomberg as at 6 July 2023
Are there merits to using a combination of Gilts and swaps when constructing the hedge?
While correlations between Gilt and swap yields are typically high, we believe that in a market environment with heightened volatility and significant Gilt issuance coupled with Quantitative Tightening (QT) there is a strong case to be made for being more dynamic when constructing a scheme’s hedge. Some schemes BlackRock works with have set a secondary objective for the LDI mandate dynamically manage the allocation between Gilts and swaps, seeking to outperform the both a Gilts- and swaps-based investable benchmark. We do this by using an asset allocation model, with forward-looking inputs. These are regularly reviewed and adjusted based on an ongoing assessment of the market with decisions made by an investment committee of seasoned investors (supported by input from across the broader BlackRock platform) with many years of market experience. Put simply, when valuation differences between swaps and gilts are high, the LDI portfolio’s allocation will aim to be in the favourable valued asset. When valuation differences between swaps and gilts are low, the allocation will be more balanced between the instruments. In Figure 4 we show how the allocation between Gilts and swaps have varied the past five years in two portfolios, one with real exposure and one with nominal exposure, where this type of strategy is deployed.
Figure 4 – Gilt positioning in two dynamically managed portfolios
The figures shown relate to past performance. Past performance is not a reliable indicator of current or future results.
Source: BlackRock as at 30 June 2023
Key takeaways for UK Pension Schemes
We would typically not expect discounting on a swaps basis to reduce the volatility of the liability valuation. While some schemes might have specific reasons for discounting their liabilities on a swaps basis, we believe that for most schemes it makes sense to continue discounting on a Gilts basis. This as we do not believe there is evidence that discounting on a swaps basis would significantly reduce liability volatility over time. Significant swaps usage in the asset portfolio also comes with its own challenges. These challenges, be they regulatory or the potential funding drag of cash collateral, are more likely to increase as the funding level improves as the swaps usage in the portfolio is likely to increase to better match the liabilities. This in contrast to a Gilts-based approach where complexity on the asset side is likely to decrease as the funding level improves and the proportion of physical Gilts held in the portfolio increases.
As schemes mature and move towards their endgame, we see some clients starting to design low dependency solutions with significant allocations to credit assets and for these schemes we believe it is worth having further discussions about aligning the discount basis with the solution strategy. These strategies come with their own challenges such as the investability of the universe that is used to construct the credit curve and the significant impact downgrades and defaults can have on the funding level. This is something we expect to cover in more detail in a future Views from the LDI Desk.
