A snapshot of the UK inflation market
Supply outlook
In March, the UK Debt Management Office (DMO) announced its borrowing remit for fiscal year 23/24, alongside the Office for Budget Responsibility’s updated forecasts for gross financing. A small downward revision of £3.3 billion was made to the remit in April, resulting in less short and long dated conventional gilts being raised via auctions but no change to inflation-linked gilts.
For fiscal year 23/24, we do not expect the same trend of fiscal year 22/23, where both syndications were skewed to the longest bond. Therefore, we have assumed a rebalance shorter in duration in our issuance projections. The decision to host the next inflation-linked gilt syndication in July, so soon after the UKTI 2045 syndication on 26 April, came as a surprise – as the bond performed poorly and there is lingering uncertainty over the pension fund community’s demand for inflation duration. In fact, minutes from May’s DMO investor consultation highlighted that “some participants did suggest that the transaction might benefit from being postponed until September or later in the financial year”.
The scheduling of a syndication in July seems to be driven more by cash management needs by the DMO rather than in specific response to end-user demand. Nonetheless, it does mean that of the £9 billion dedicated to inflation-linked gilt syndications, at least £7.5 billion could already be allocated barely four months into the fiscal year. That said, the DMO has pointed to the flexibility afforded to it by the £12 billion of unallocated gilt sales, which could supplement the size of the final inflation-linked gilt syndication later in the fiscal year, or potentially even provide an inflation-linked gilt tender if demand is strong.
In the meantime, attention turns to the choice of bond for July’s syndication, with the following tipped as possible candidates:
- A second tranche of the UKTI 2045 – this was cited as the most likely candidate, given the large unfilled switch book and low allocations at the syndication in April. This would also reduce the potential for the market to be saturated with duration, particularly if the issue size is only £3 billion.
- A new 30-year bond – a bond maturing in 2054 would fill the maturity gap well and generate switch interest, whilst creating a future 30-year benchmark bond. We think this is more likely due next fiscal year as the UKTI 2051 is not yet at benchmark size and is still being tapped via auction. This would also supply more duration, which would be more problematic for the market to digest whilst end user demand remains tepid.
In our issuance projections for the fiscal year, we assume our base case scenario of the UKTI 2045 for July’s syndication, followed by the UKTI 2073 for the remaining syndication. Despite the shorter average maturity in syndication candidates this fiscal year, we still expect a rise in inflation duration issued compared to the previous fiscal year simply due to the larger amount allocated to inflation-linked gilts.
Figure 1: Inflation-linked issuance1
Source: Columbia Threadneedle Investments, DMO, as at May 2023
Figure 2: Distribution of inflation exposure issued, by maturity buckets
Source: Columbia Threadneedle Investments, DMO, as at May 2023
Market liquidity
Market transaction volumes have picked up compared to when we last polled our counterparties – liquidity is improving but has yet to recover to pre-gilt crisis levels. The Financial Conduct Authority and the Pensions Regulator released their recommendations for LDI managers at the end of April, which were as expected. This could be one of the reasons for the (tentative) return of LDI buying at the start of May, particularly at longer maturities. It remains to be seen whether hedging activity will return, as some schemes may have been awaiting these publications before proceeding with de-risking. The last weeks of May saw a gilt sell-off, where despite a circa 1% rise in 30-year real yields compared to the turn of the first quarter, there was little evidence of any meaningful selling. Bid-offers widened out over this period as market makers were averse to holding on to risk positions for too long.
At the 10, 30 and 50-year tenor points respectively, mean bid-offer spreads2 were 1.2bps, 1.1bps and 1.3bps for RPI swaps; and 1.2bps, 1.0bps and 1.3bps for index-linked gilts.
Figure 3: Liquidity tracker based on a poll of investment bank trading desks’ bid-offer spreads for RPI swaps and inflation-linked gilts (intra-day) at 10-year, 30-year and 50-year assuming trading £50k risk (IE01)
Source: Columbia Threadneedle Investments
Medians are based at 100% of bid-offer spreads in February 2023. The movements in median, from 100%, indicate outright changes in transaction costs, while the changes in the upper and lower quartiles indicate the dispersion of these costs.
Compared to February, transactions costs in the three months to end-May fell for both RPI swaps and inflation-linked gilts, and both instruments also experienced larger cost dispersion across all maturities (except 10-year gilts which stayed broadly unchanged).
Gilt versus swap inflation
The main drivers of the differential between gilt and swap inflation over the three months to end of May have been the nominal curve, supply outlook and the expected strong bulk purchase annuity (BPA) deal pipeline. Following April’s remit revision, the high levels of conventional gilt supply continued to cheapen nominal spreads, particularly at longer maturities, which in turn, weighed on gilt inflation versus swap inflation.
The last week of May was a notable period as we saw a sharp underperformance in the gilt market – the nominal yield-led sell-off pulled real yields higher with it. An upside surprise in the inflation data pushed nominal yields and gilt inflation higher. Long-end inflation-linked gilts were hit the hardest, breaking through the significant 1% real yield barrier, and at one point looked to be headed towards the record highs of the September / October gilt crisis. It is worth noting the dissimilarities, however, between May and then – given the orderly nature of the sell-off, reduced leverage and more stringent rules in place for collateral buffers, a rapid runaway sell-off seemed unlikely this time. The front end of the curve led the yield move; hence curves bear flattened (economic contraction), rather than bear steepened (rising inflation expectations) like last time.
Another factor driving the cheapening of the real yield curve (in particular at longer maturities) has been the ongoing inflation-linked gilt asset-swap selling, which is suggestive of bulk purchase annuity (BPA) deals being hedged in the market. With the significant rise in gilt yields over September / October, funding levels improved and hence the flurry of BPA deals announced since is unsurprising. Swap inflation has outperformed gilt inflation year-to-date, particularly at the 20 to 30-year tenors, and this trend is set to continue if more deals are brought to market. That said, trading volumes remain relatively light and the differential between gilt and swap inflation has more recently been driven by the gilt leg of the trade (and inflation-linked gilt issuance).
Figure 4: Relative z-spread for generic inflation-linked bonds versus comparator SONIA z-spread (3 months to 31 May 2023 highlighted), where higher (lower) level indicates swap inflation outperforming (underperforming) gilt inflation
Source: Barclays Live
CPI market update
The fifth allocation round of the Government’s Contracts for Difference (CFD) scheme closed for final submissions at the end of April. This is the first annual CFD (previously biennial), for which results are to be published around the start of July at the earliest. This usually results in significant corporate supply coming to market, as the winning bidders hedge their CPI-linked income streams. In terms of timescales for hedging, there is often a lag following the announcement of results, although larger firms may consider some initial hedging before moving this into a financing vehicle later.
The RPI-CPI forward wedge as at the end of May compared to the end of February has not been re-marked significantly. The divergence in CPI and CPIH has previously been cited as one of the possible reasons for a non-zero post-reform forward RPI-CPI wedge. Between CPI and CPIH, the difference arises from owner occupiers’ housing costs and council tax. Housing-related factors are linked to domestically generated inflation rather than imported inflation, so CPI is more sensitive to the latter (i.e. it comprises more import-intensive components). With import-intensive components currently contributing more to CPI than they have historically, softer CPI figures going forward should lead to the CPI-CPIH wedge (and hence the RPI-CPI wedge) contracting from current levels to trend toward zero.
Figure 5: Indicative spread between RPI and CPI swaps expressed as a strip of forwards3, at 25 November 2020 (RPI reform announcement), 28 February 2023 and 31 May 2023
Source: Columbia Threadneedle Investments, Morgan Stanley