Physical gilts have generally offered a yield pick-up over equivalent swap-based exposure. Yet this premium has declined over the past two years as pension funds have continued to invest heavily in gilts to match liabilities. Let's consider how much this premium might have to fall for investors to swap bonds for swaps.
I was on holiday a couple of months back. Whilst I could just about stomach prices in half-term, I have grave concerns about the summer holidays! Maybe I’ll be selling the alternative 'tent in garden' option to the family instead… And with that tenuous link in mind: at what point will pension funds find gilts too expensive and try swaps for size?
Many a commentator has spoken over the years about how expensive gilts are looking, yet pension funds continue to increase their hedging in order to manage their risks. Of course, where investors do not have long-dated liabilities, they can look elsewhere, which is just what our Multi-Asset team did in order to gain some inflation exposure.
Nonetheless, for a genuine balance sheet or funding level hedge, pension funds are essentially limited to investing in either gilts or swaps. So, what do we need for pension fund buying to shift from gilts to swaps? To answer that we need consider how pension funds will weigh up the prospective risks and returns.
As the chart below demonstrates, gilts still offer a premium relative to swaps, albeit the trend in recent years and the last few months has been for that premium to reduce. The exact premium varies by bond maturity and whether they are fixed or inflation-linked (for example, the additional points on the chart show a 10-year gilt and a 50-year index-linked gilt) but for the purposes of this piece we'll stick to 50-year gilts.
For a leveraged investor, that premium has arguably fallen quite close to zero at times if some conservatism is built into a long-term funding assumption (the long-term cost of the borrowing required to invest with leverage).
Without wishing to go into all the detail here let’s take SONIA +10 to 20bps (0.1% to 0.2%) as a standard funding assumption (SONIA is a UK interest rate and stands for the Sterling Overnight Index Average) but add on some buffer for conditions in times of stress.
We estimate that gilt yields should be at least 20bps to 30bps greater than swap rates
That gets us to around 30bps which we remove from the blue line to get to the yellow line (swaps can be thought of as funding at SONIA +0bps). For those interested in what has happened to repo (repurchase agreement) markets in stressed times the answer is in our LGIM LDI Monthly Wrap (spoiler alert - the answer is around SONIA +50bps for a short period).
And yet would the flows into gilts really stop at these levels? My view is ‘no’ as those buying gilts without leverage are still likely to be attracted to gilts as they don't have the same funding concerns. Put another way, the alternative approach of swaps fully backed by cash is not an obvious solution.
We should also consider the risk side of the equation. The typical method of matching liabilities basis remains gilts-based and moving from gilts to swaps (especially SONIA swaps) sounds like a pretty big exercise to me.
But I wouldn’t be surprised in years to come to see gilts yielding only 10bps more than swaps
On the other hand, those pension funds moving towards a more credit-based discount rate may be able to deal with this but there will inevitably be some inertia as a result of the complexity.
So when will pension funds swap gilts for swaps? Well, I see some pension funds stopping their purchase of gilts at spreads of 20bps to 30bps over swaps thereby providing some resistance. Nonetheless due to the additional complexities on the liability side of moving into swaps as well as demand from unleveraged schemes, I wouldn't be surprised in the years to come to see gilts yielding only 10bps more than swaps.
Index management is about much more than benchmark replication. Clients aware of this fact are quickly shifting the way they evaluate funds in search of something that exceeds traditional expectations.