The Retail Note - Retail sales: consumer storm before the lull?

Key Messages

• Stellar retail sales performance in Sep
• YoY retail sales values +4.0%, volumes +2.3%
• Implied shop price inflation of 1.7% lower than headline inflation of 3.8%
• MoM volume growth +0.5% much higher than economist consensus (-0.2%)
• Q3 retail sales values +3.2% an acceleration on Q2 (+2.1%)
• Likewise Q3 retail sales volumes (+1.5% vs +0.7%)
• Grocery sales surprisingly weak (vals +2.7%, vols -0.7%)
• Non-food sales growth far stronger (vals +3.8%, vols +3.1%)
• Charity Shops, Electricals, Furniture standout categories
• Strong performance from Clothing (vals +7.8%, vols +6.6%)
• Challenging month for Textiles, Carpets, PCs & Telecomms
• Online sales grow +1.4% MoM and +5.6% YoY
• Online penetration increases by +20bps to 28.0%
• The Autumn Budget looms large…
• …but will not necessarily destabilise retail spending.

The forthcoming Autumn Budget is all-pervasive. The narrative around today’s retail sales figures was always going to go down one of two paths. “Weak, with consumers tightening their belts in anticipation of a whole host of nasties that will hit their pockets come 26 November”. Or “Strong, with consumers having a last hurrah ahead of a whole host of nasties that will hit their pockets come 26 November”.

Thankfully, the latter of these has come to pass, with retail spending generally (but not universally) strong in September. But the secondary narrative to this positive outcome is inevitably “it can’t last”, when actually, just maybe, it can.

The credibility of the ONS retail sales data has come into question in the wake of the delayed release of the July 2025 figures and subsequent revision of the year-to-date numbers. While some of the numbers may have changed (as I will go on to explore), the direction of travel is relatively clear – and generally positive.

The headline stats

Putting the widely-quoted but largely meaningless month-on-month and quarter-on-quarter figures to one side, the year-on-year statistics were exceptionally strong. In September, retail sales values (exc fuel) were up +4.0%, with volumes ahead by +2.3%. The revisions to historic data have muddied the waters slightly in terms of context, but broadly this was the best monthly value performance since the inflation-fuelled days of 2023. In volume terms, this was the third best monthly performance since COVID (after April 2025 and December 2024 – more comment on that to come).

Equally positive is the implied level of shop price inflation of +1.7%. While the nation heaved a sigh of collective relief as headline inflation stayed at +3.8% earlier in the week, retail even had the decency to stay below the government target of +2.0%.

The quarterly picture is perhaps even more edifying. Q3 retail sales values (exc fuel) grew by +3.2% year-on-year, with volumes up by +1.5% (and implied shop price inflation of +1.7%). Strong in their own right, the bigger picture is of accelerating value and volume growth as the year has unfolded (Q1 vals. +1.4%, vols. +0.1%, Q2 vals. +2.1%, vols. +0.7%). Strong value and volume growth, an improving monthly and quarterly trend, coupled with manageable and stable inflation – could things actually be much better?

Of course, there is scant mention of any of these figures in the ONS release itself, nor by extension in any of the reporting thereof. For what they are worth (precious little), the reported month-on-month and quarter-on-quarter figures also had a veneer of strength. QoQ volumes were up +0.9%, while MoM volumes were up +0.5%. The latter figure completely defied economist consensus forecasts of -0.2%. Rather than just crying “it won’t last”, maybe they should be admitting that they got it spectacularly wrong, yet again?

The great rebase

The delay in the August release and subsequent revision of the data has undermined the credibility of some of the data. To be fair, I have been pretty vociferous in the past in my criticism of certain aspects of the ONS retail sales data, particularly their attempts to “seasonally-adjust” the raw data in a vain effort to bring retail sales into line with other economic metrics that are not subject to seasonal swings (particularly GDP).

Making adjustments on the basis of the length of reporting periods (e.g. 4 vs 5 weeks) is understandable, but trying to “smooth out” seasonal event such as Christmas and Easter is a total nonsense. A situation whereby January’s MoM retail sales figures are up bears absolutely no relation to reality. The solution? Report year-on-year figures, as I always advocate. That way, you are nearly always comparing like-for-like. The only real exception is Easter, which may fall in either March or April. The solution here: report two months combined, again year-on-year.

Without going into massive detail, the rebase of the retail sales figures was essentially the ONS revisiting the way it has historically “seasonally-adjusted” the data. None of the “non-seasonally adjusted data” has been revised in any way.

What the revision has meant, in very simplistic terms, is that the figures for the year-to-date (January to July) were marked down slightly i.e. retail spend was not as strong as originally reported. In general terms, the downward revision was more pronounced in food rather than in non food.

In monthly terms, the suspiciously strong January figures were indeed discovered to be suspicious and were subject to significant downgrade (+1.5% to -0.1%). Nor was the bonanza that was April quite so bonanzary (+6.1% to +3.9%) and there were also minor downgrades to the figures for February, May and June, but an upwards revision in March.

To my eyes, two things stand out. The grocery figures still do not look credible. Year-to-date grocery retail sales values have grown by just +1.3%, but volumes have declined ca. -1.5% to -2% and remain in negative territory. This stands in sharp contrast to what the major players (in a highly concentrated market) are saying.

Secondly, the revisions now reveal that last Christmas was far, far stronger than was originally reported. Retail sales values for December have been upgraded from +2.6% to +3.6% and volumes from +2.0% to +2.6%. Of course, no one cares about that now as it is yesterday’s news. But it does go some way to explaining the mismatch between the then dismal retail sales figures and the general upbeat reporting from the retailers themselves. And the revised figures do create a more demanding benchmark this time around, so expect the cycle of festive doom mongering to continue…

Performance by sector

 The usual disparities between individual sub-sectors. Curiously, Computer & Telecommunications retailers were singled out by the ONS as a top performing category, again underlining the folly of month-on-month reporting. On an annualised basis, Computers & Telecommications saw sales down -5.7%, with volumes down -1.6% (and implied deflation of -4.1%). Hardly a strong performer.

All non-food saw value growth of +3.8%, with volumes up a very healthy +3.1%. Perhaps surprisingly, Charity Shops were the top performing category (vals. +31.9%, vols 31.5%), followed by Medical Goods (+28.9%, +25.1%) and Electricals (+9.9%, +12.1%). Clothing continued its strong run (+7.8%, +6.6%) and enjoyed its best quarterly growth since COVID (+7.8%, +6.8%).

At the other end of the performance spectrum, another very challenging month for Textiles (values -33.7, volumes -34.4%) and Carpets (-7.8%, -9.7%).

Online spend grew by +1.4% MoM and +5.6% YoY in September (+3.5% QoQ, +5.0% YoY). In both cases, this was higher than overall retail spend (+0.7%, +4.0%), so online penetration rose from 27.8% in August to 28.0% in September.

For more detail please refer to the accompanying Retail Sales Dashboard.

Wider context

When there is a strong set of retail sales figures, the doom mongers tend to seek out consumer confidence metrics to support their cause. But even they aren’t supplying them with the ammunition they need, Gfk’s figures released today showing a marginal positive uptick in underlying consumer confidence.

Can it last, or is this just a final fling before an inevitable downturn? Obviously, the Autumn Budget looms large and speculation will only intensify in the coming weeks as to what may or may not happen as a result of this (watch this space, or rather the Knight Frank Insight website). Last year’s Budget was brutal for the retail sector, but predominantly affecting retail operators rather than consumers themselves. That could yet be the big difference between this year’s and last’s.

But how wedded are retail sales to politics and, dare I say, economics generally? Did retail sales fall off a cliff in the wake of the disastrous Truss mini-Budget in September 2022? No, they didn’t, they maintained a fairly steady pattern of monthly growth (ca. +3.0%) over the months that followed. We are still blind as to what might happen in November, but it is premature to assume that retail sales will be negatively impacted.

And, of course, retail is now approaching its peak season (the desperately unimaginatively-titled Golden Quarter). Some research houses have already put out their festive trading predictions (those that are predicting very low value growth and a volume decline would do well to study the ONS data). Ours will follow in the coming weeks. Spoiler alert – neither boom nor bust. Decent not disastrous. With volume growth. But caveated to the hilt with warnings of data misrepresentations.

And yes. Recent retail sales performance can continue.