LDI Key Talking Points – August 2020, Part 2

With the consultation on RPI reform closing on 21^st August this, second, instalment of LDI talking points for August focuses on the inflation market. In particular we focus on estimating how much the market may be pricing any upcoming reform as well as looking in more detail at the most recent inflation figures, which came in significantly above expectations.

Quantifying the market implied levels of RPI Reform

As the RPI reform consultation draws to a close, it seems appropriate to attempt to answer the question of how much of any likely outcome has been priced into current market levels. However, as with any question of quantifying the direct market impact of one particular factor against a backdrop of other variables, there are some very strong and contentious assumptions required.

Before we get to assessing what proportion of any potential reform is priced in, we first need to estimate what the long term RPI-CPIH spread could be. This can spark significant debate, with a wide range of estimates and challenges around using a simple historic average due to the inclusion of mortgage costs in RPI over a period where rates have consistently fallen imposing a downward bias to both RPI and the RPI-CPIH spread. There are many defensible assumptions which can be taken, for example using the 2015 OBR estimate of a 1% spread for RPI-CPI as a base. However, for the basis of this analysis we are using an assumed spread of 85bps, in line with our previous analysis on RPI reform.

It is also worth noting that we have made several other simplifications, which will also have an impact on outcomes, such as using the prevailing spot starting swaps for our analysis. Whilst a 10y swap out of 2019, when reform discussions began, would have matured before the 2030 ‘deadline’ for implementation, a 10y swap out of today would mature in 2030. There would also be more natural ‘rolldown’ from swaps shortening over time. As such, there are certainly ways in which the analysis can be refined, but given the range of other simplifications required, the approach below should still provide some initial insights.

We can compare four separate simplistic quantification approaches:

Approach 1 : Long end RPI vs comparators (Euro HICP, US CPI)

This approach looks at comparisons between RPI forwards and Euro HICP and US CPI. The significant assumption required would be that there are no other factors driving any deviation between RPI and HICP/US CPI aside from RPI reform. Clearly over short horizons, this is a less contentious assumption, as if the time period encompasses significant announcements on RPI reform, it is likely to be a dominant factor. However, over a long horizon where we’re seeing factors such as Covid, Brexit, Oil price moves etc, the information value generated from this is more limited.

Comparing the relative move in the 3 curves since the Lords report in Jan-19 shows that the UK RPI curve has inverted an additional 29.3bp compared to average levels across US CPI and Euro HICP

Approach 2: Quantifying moves immediately after reform announcements

Alternatively, we can look at the moves immediately after ‘big’ reform announcements. By focusing on these days only, the data is cleaner and less exposed to the noise of other market drivers. However, this neglects the longer term impacts that are felt from the market ‘pricing in’ changes over a longer period rather than the immediate move post-headlines.

We have focused on two events. The Lords publication in Jan19 and the reaction to the letter from then-Chancellor Sajid Javid in Sep19. We look at summing the market moves across the days of these two events.

We can see moves of ~25bp at the long end for the combination of the two days, although it’s also notable that the very front end also dropped, despite reform being stated to begin at 2025 at the earliest. This suggests either that there were other factors at play, or that a general weakness in RPI across the curve also had an impact.

Approach 3: CPI – RPI swap spread

Whilst CPIH doesn’t trade in the market, CPI has transacted in a series of larger transactions, albeit in bursts of sporadic activity, with total volumes being a tiny fraction of the RPI market. Historically, insurance interest driven by capital relief led to CPI trading at a premium to what many assumed as ‘fair value’. However, as increased supply arrived, predominantly from EMR (Electricity Market Reform) projects, CPI swap inflation levels available reached  a level many considered approaching a long-term neutral spread. In early September-19, prior to Javid’s letter, market levels reflected this balance. Whilst in the few weeks following this, the RPI-CPI spread in swaps tightened, the moves were still relatively marginal. However, the gap between RPI and CPI swaps has continued to fall, with Aug-20 wedge levels implied by swaps less than half of what they were. It is worth noting that over the period of RPI reform discussions, the already comparatively low volumes of CPI supply have diminished further.

To some extent this can be considered a ‘clean’ way of looking at RPI reform pricing, given it does not require as many assumptions around exogenous factors. However, due to the thin CPI market, there are clear supply and demand factors at play, and it is contentious to base a conclusion on an instrument that is not trading with any meaningful volume, at present.

10y10y RPI-CPI forward spread was 76bp on 03-Sep-19 and 21bp on 14-Aug-20, a tightening of 55bp.

Approach 4: RPI move vs historical levels

As the most simplistic approach, we can look at the long-term levels of RPI in forward space. The assumption would be that if there is a risk premium and/or supply and demand driver that leads to a premium above the long run fair value level, we could assume that these disconnections would stay constant over time. Obviously, this is a contentious approach, particularly over longer time horizons. There is some similarity between this approach and the cross-market comparison, however rather than taking other inflation curves as a benchmark, we are taking a historical comparison from the same curve.

The 1y forward in 20y has fallen from 3.499% on 16-Jan-19 to 2.65% on 18-Aug-20. If we were assuming moves in the forwards beyond 2030 are only driven by RPI reform, we would expect a parallel shift. However, whilst the drop has been 85bp in 20y, it is 35bp in 10y and 70bp in 30y. The inconsistency across long tenors highlights the weakness in this approach. There may be some justification in averaging across tenors to smooth out some noise. The average change across the 1y forwards between 10 and 30y is 70bp.

To summarise, getting a conclusive view on how much has been priced in is challenging, and requires assumptions in both what the “neutral” level of spread might be, as well as strong assumptions underpinning flawed estimation approaches. However, looking at a range of approaches can give a broad sense of what may be reflected in market valuations. Based on the summary table below, it would not be unreasonable to assume somewhere in the region of half of the impact of a change has been priced beyond 2030, with a significant margin for error around this estimation.

With a lot of uncertainty on how much is in the price and the outlook for inflation (see below) it is a difficult call for pension schemes to make definitive changes to their strategic inflation hedge. We would favour using the ability to tactically take advantage of where pricing on the curve may look stretched, for example, is enough probability given to reform actually starting from 2025? Is premium for 15yr inflation relative to 30yr inflation shown on the chart in approach 4 reasonable? We do not expect a concrete update from the Treasury soon, with the timeline continuing to be “in the Autumn” for now. Liquidity, supply and price moves in the inflation market, are likely to reflect this continued uncertainty.

Above expectation inflation print for July

The UK inflation numbers for July were released on August 19th. CPI came in at 1.0% YoY vs a market consensus of 0.6%, whilst RPI came in at 1.6% YoY, also 0.4% ahead of predictions. Whilst such a strong print, at first glance, could be taken alongside the ongoing expansionary fiscal and monetary policy changes as a sign that inflationary pressure is building up in the system, a look below the surface shows that a large component of this is optical, with many of the risks ahead skewed towards lower inflation.

Some of the largest impacts came through the clothing sub-index, where normally we see very strong seasonal patterns with regular annual sales, in the summer and in Dec/Jan. However, this year, we saw clothing prices atypically cut in March and April as retailers tried to stoke demand against an uncertain backdrop. This meant lower YoY clothing inflation numbers in the last few months, which has been unwound for the July figure, where the lack of an incremental discount in July has boosted YoY inflation rates.

Looking ahead, we have the impact of the VAT cut to come from next months data onwards, and a lowering of energy bills confirmed by Ofgem from October onwards. We will also see the impact of the Eat Out to Help Out (EOHO) scheme. This is being reflected using the formula: EOHO price = (Thu to Sun weight) * Full price menu + (Mon to Wed weight ) * 50% * Full Price menu. By removing the £10 cap, this will lead to an even larger reduction in prices.

Also, it is worth noting that, due to difficulty collecting data over the pandemic over the past months, the ONS has been using more imputed prices and online sampling. They are now reverting back to more traditional sampling methods, but we will likely see noise in the data over the period of adjustment.

So, to conclude, we don’t think that the latest print should be taken as a sign of more prolonged inflation, and we would expect lower, if volatile, prints in the months ahead.

The opinions expressed are as of August 2020 and are subject to change at any time due to changes in market or economic conditions. The above descriptions are meant to be illustrative. There is no guarantee that any forecasts made will come to pass.