The Long And Short Of The Gilt Market

In recent years the gilt yield curve has maintained a particular shape, with long yields inverted. Alan Cubbon argues this is likely to continue through the expansion of the market necessitated by the coronavirus, and therefore that gilt investors who have no particular need for long-dated bonds should consider excluding them from their benchmarks.

In March the UK Chancellor announced a £330 billion loan scheme designed to help businesses stay afloat. This was the largest in a series of fiscal responses to the coronavirus pandemic whose total cost to the country the think-tank Bruegel puts at £476 billion as of the end of May . With the economy facing a deep recession the ability of the government to recoup these costs through taxation is likely to be extremely limited, at least in the short term, so an increase in borrowing through greater gilt issuance would seem to be the only option.

According to the UK’s Debt Management Office, as of the end of February, just before the current crisis began, the face value of all outstanding conventional and index-linked gilts totalled £1.64 trillion. As of the end of May this had increased 4.7% to £1.72 trillion. An extra £476 million on top of the pre-crisis figure would amount to an expansion of 29%. That leaves a further expansion of 23.2% from the May figure to come. Can the gilt market absorb this extra issuance? With yields close to record lows and the prospect of untold levels of Quantitative Easing it feels hard to worry that the market will become saturated, at least assuming the new issuance is phased in over a reasonable time-frame.

Might there, however, be a more subtle effect on the characteristics of the market: in particular the conventional gilt yield curve? In recent years the level of the curve has varied, but it has always retained a characteristic shape, as can be seen in the charts below.

Selected gilt yield curves, 2018 to today

Source: Tradeweb
(Tradeweb is responsible for calculating Gilt and Treasury bill end-of-day reference prices based on data from the Tradeweb dealer-to-client UK Gilt trading platform. Tradeweb derives end-of-day UK gilt strip prices from a ze-ro- coupon yield curve calculated from the reference prices of conventional gilts. This approach follows the practice of the DMO and is in accordance with Bank of England publications.)

In each chart – and in fact the pattern holds for several years earlier – past the shortest maturities the curve quickly becomes upward-sloping up to about 25 years, whereupon yields then begin to fall. The reason for this fall in yields is that long gilts are eagerly snapped up by insurance companies and pension funds. Such firms together make up the largest holders of gilts, owning an estimated 33% of the market . They have long-term liabilities whose repayment profiles they seek to match using long-term, low-risk cashflows that these long gilts provide. But could the increase in gilt issuance precipitate the end of this long-maturity inversion?

The outstanding face value of conventional, fixed-rate, gilts is £1.28 trillion or about 74% of the total: the remaining 26% being made up of index-linked gilts. (This proportion has remained pretty steady over recent years.) £264 billion-worth of conventional gilts, or a little shy of 21% of the market, have a maturity longer than 25 years, which is close to where the yield curve peaks. With the current longest gilt maturing in 2071, this means that the longer half of the curve is much less densely-populated than the shorter end, as is also shown in the chart below.

Source: Debt Management Office, 30th May 2020

Consider next the extreme long end of the curve. In 2012 the UK Debt Management Office conducted a consultation on the market’s appetite for “Super-Long and Perpetual Gilts”, where a “super-long” gilt was defined as one with a maturity of at least 50 years. (https://www.dmo.gov.uk/media/14574/cons20121205.pdf). The consensus among replies was that there was a healthy appetite for gilts up to 60 years, but little desire for maturities above that, or for perpetuals. It was noted, not unconnectedly, that pension fund liabilities tail off rapidly beyond 60 years. In terms of supply, it was thought that an issuance amount of between £5 billion and £10 billion of super-long gilts each year would strike a balance between maintaining liquidity without saturating the market. Following the consultation the DMO raised the maximum maturity cap of 50 years slightly, issuing a first 55-year bond in June 2013. Today, there is a single gilt with a maturity longer than 50 years: the aforementioned 2071 bond with a face value of £14 billion.

These amounts are small compared to the overall increase in debt, but the consultation suggests that the super-long sector could cope with a quick doubling in size, and more later, rather than just the 29% increase for the whole market envisioned earlier. If this whole market increase boiled down to basically scaling up the pre-crisis maturity profile, then there’s little to suggest that the sparsely populated long and super-long parts of the curve would be unable to cope. In any case, if certain long maturities were to approach saturation and yields began rising, it seems unlikely that the DMO would continue to issue into that part of the curve, where they themselves would be responsible for pushing up their own financing costs.

Perhaps a greater long term threat to the status quo comes from abroad. Other countries are feeling pressure to increase their bond issuance. If, for example, the US Treasury were to increase duration at the same time, that could offer an alternative to gilts for UK insurers and pension funds once the foreign currency exposure is hedged out. In the long term tax revenues could also have scope to increase and business loans be repaid.

In the nearer term though, provided that the maturity profile of new issuance is broadly in line with recent history and the rate of issuance is not too great, it appears likely that we will continue to see the familiar shape of the gilt curve for a while longer.

A consequence of this should be that non-insurer and pension fund investors who use full gilt-market benchmarks take a moment to consider if this is ideal. For them the curve beyond 25 years will remain relatively unappealing. Why fight to purchase expensive longer bonds if they have no real need? Would they instead be better off restricting themselves to shorter gilts of less than 25 years maturity, at least in their benchmark? Of course, super-long bonds do have some nice traits: their high convexity being one. But this is, literally, a second order consideration and the purchase of such bonds could best be left to an active manager placing tactical bets. Aside from this the long, low-yield sector of the gilt market could be left to the large insurers and pension funds for whom it is especially attractive.