The UK’s inflation figures were wrong and now we just have to live with that

It’s been two lengthy weeks since we final wrote concerning the Workplace for Nationwide Statistics.

Fortunately, the stats gang have finished some posting of their very own within the meantime, providing an replace on efforts to combine groceries scanner knowledge into Britain’s inflation figures.

For those who’ve been dwelling beneath a rock and missed out on this main information, the ONS has been making an attempt — for a while — to impute costs gathered from scanners as a way to observe quite a lot of costs, moderately than their conventional system of observing less of a lot of prices.

The opposite good thing about such a system is that it helps statisticians higher perceive the completely different portions of comparable merchandise being purchased, permitting them to weight gadgets extra appropriately throughout the total index calculation.

The opposite, different profit is that this: scanner knowledge would take into consideration costs truly paid, moderately than the present system of monitoring listed.

To know why this issues, think about the Tesco Clubcard. As we just lately famous, Britain’s largest grocery store chain is a big deal relating to nationwide statistics. And as any common Tesco shopper is aware of, having its loyalty card — the Clubcard — is more or less obligatory to unlock its best offers (caveats).

In keeping with Kantar, Tesco currently has a 27.8 per cent share of the UK’s grocery market. In keeping with Tesco, Clubcard gross sales penetration within the UK is 84 per cent. Ergo, just below 1 / 4 of all UK grocery spending is completed at Clubcard costs, and people costs are massively vital if you wish to measure the actual charge of inflation.

However the ONS ignores Clubcard costs. Right here’s an extract from its Client Costs Indices Technical handbook, the inflation bible:

Discounted and subsidised costs are solely recorded if out there to anybody with no circumstances of sale, in any other case the non-discounted or unsubsidised value is recorded. Cash-off coupons and loyalty playing cards are excluded.

The logic right here makes some sense. Because the ONS places its elsewhere, “our primary assortment apply is to solely think about reductions if they’re out there to all”. And naturally Clubcard costs are solely out there to the good many individuals who be a part of that membership (principally buying and selling their knowledge for reductions).

However when it comes to capturing inflation when it comes to the change in how a lot persons are truly paying for items, it’s a multitude.

The ONS was requested about this in a freedom of data request summer season earlier than final. In response, they mentioned:

[At] the top of 2022 we began accumulating loyalty card costs alongside the costs used for dwell manufacturing of our shopper value statistics. We just lately used this knowledge to conduct a casual (inside) evaluation to evaluate the impact loyalty card reductions would have on our inflation measures in the event that they had been included. This train didn’t produce proof {that a} hypothetical index created by utilizing the low cost costs can be materially completely different to revealed CPI on the headline or divisional stage.

In brief, they discovered Clubcard costs don’t matter.

The scanner knowledge was alleged to be launched in March this 12 months, however that is the ONS we’re speaking about, so because reasons they’re now scheduled to land in March subsequent 12 months.

The ONS provided some preliminary indications of the change’s influence in January (written up by Chris Giles here). And this morning, it provided slightly extra.

Now we have produced indicative estimates of the influence of introducing groceries scanner knowledge into UK shopper value statistics; estimates are topic to vary following additional high quality assurance.

So says “Transformation of UK consumer price statistics, groceries scanner data analysis: April 2025”. It provides:

The typical indicative change to the Client Costs Index (CPI) annual charge from the introduction of groceries scanner knowledge between January 2019 and June 2024 was detrimental 0.04 share factors; this already adjusts for the adjustments we launched in February 2025.

Now, 0.04 share factors doesn’t look like a lot. And, certain, usually the distinction appears to have been fairly small total throughout that interval.

However we’ve to take problem with this.

Firstly, exhibiting an absolute change over a protracted time frame appears certain to provide a reasonably small quantity, given you may anticipate changes to happen in each instructions. Additionally, a 2019–2024 window barely dilutes the entire intense inflation wave (with its bizarre dynamics) we had in the course of the interval studied.

Secondly — because the ONS knowledge exhibits however doesn’t actually dwell upon — even when the general common change is small, a few of the adjustments have been fairly large.

Right here’s a chart from the ONS launch:

The model of their article is responsive, so it’s not laborious to see the exact knowledge on-site, however to us this nonetheless qualifies as a licensed Axis of Evil™. By exhibiting the 2 compounded 12-month charges (the traces) on the identical axis because the unfold between them (the columns), you make these columns look small.

And as one other chart additional down the article exhibits, these columns should really feel large:

Headline CPI would have been 1 / 4 of a share level decrease in March 2024 utilizing this knowledge. Let’s forged our minds again to how mainFT covered that month’s CPI figure at the time:

Client costs rose at an annual charge of three.2 per cent in March, down from 3.4 per cent in February, the Workplace for Nationwide Statistics mentioned on Wednesday. The determine was barely increased than the three.1 per cent forecast by economists polled by Reuters and the Financial institution of England and above the BoE’s 2 per cent goal. 

Merchants in swaps markets at the moment are betting that the BoE will start decreasing its benchmark charge from a 16-year excessive of 5.25 per cent in both September or November, having totally priced in a lower for September earlier than the ONS launch.

With its “improved strategies”, the ONS now thinks the year-on-year change in CPI that March was 3.1 per cent. Alter for the scanners, and that turns into 2.8 per cent — 0.4pp beneath the originally-reported determine.

That’s a big swing! It is a large deal! It’s clearly materially completely different*, and — assuming it’s truly more true to the fact of inflation — it stays materially vital that the ONS is taking so lengthy to get this improved methodology in place.

As FTAV Friday charts quiz T-shirt owner Gregory Boggis factors out on X:

Let’s simply be glad the UK by no means makes major spending decisions primarily based on a couple of billion kilos of fiscal swing.

* OK, OK, clearly it wouldn’t have mattered as a result of markets are completely environment friendly and due to this fact have already priced in any and all errors which may happen within the ONS’s CPI calculations.