MNI INTERVIEW2: Scanner Data Dampens Volatility- ONS's Benford

The UK’s official statistic’s body’s planned move towards using more checkout scanning data rather than physically collecting prices from retailers looks likely to provide less volatile inflation numbers, senior Office for National Statistics official James Benford told MNI.

The ONS plans to replace physical price collection for half the grocery market by March 2027, which should capture the effect of the widespread use of loyalty cards by retailers on the prices consumers actually pay, which are often discounted versus the display price, Benford said in an interview.

"I don't see why it would be deflationary or inflationary in terms of the average effect, and that's broadly what we find by looking at the average over the period where we have both scanner and sampled data in the past," he said, noting that retailers which discounted too persistently would run into trouble.

“There are some periods where various variations of discounting push up a bit and there are some periods where it pushes down," he said. "A period in 2024 where there was heavier discounting, was a period when food prices were going up ... the discounting was acting to reduce volatility, rather than having a persistent effect in one direction.”

 IMPACT SUBDUED

 At present around a third of the CPIH inflation measure comes "from more comprehensive data sources,” including scanner data and directly-collected rental data, rather than the traditional sampling of quoted prices.

 "On average over the last five to six years. If you look at the impact of the previous methods on consumer prices relative to the new ones, there isn't much of a differential. I think it's less than point one on average," Benford said.

 "There are periods where there was a bigger impact. And as I mentioned, when food price inflation was high, there was a bigger drag, suggesting heavier discounting.”

 The maximum effect of the scanner data was 0.2 to 0.3 percentage point on headline CPIH but "on average, over a longer period, the impact on inflation has been much smaller,” he said. (See MNI INTERVIEW: Inflation Expectations Resisted BOE Hikes-Weale)

 NOT HARMING LINKERS

The BOE, which is mandated to assess whether any change in inflation data has a materially detrimental effect on pricing for inflation-linked gilts, has already looked at the impact of the increasing use of scanner data, Benford said.

"They viewed it to be a significant change to the index, an improvement to the index, but they didn't see any material detriment. Which is one way of inferring that they don't see a deflationary impact, otherwise they might come to a different judgement.”

MONTHLY GDP

The ONS started highlighting three-month-on-three-month rolling growth rates in its September GDP bulletin, and Benford said this came in response to a tendency by journalists and other observers to over-emphasise the importance of small variations in monthly GDP data, in particular if these show a shift from expansion to contraction or vice versa. (See MNI: BOE Needs More Varied Scenarios - Bean, McMahon)

"The monthly data is useful to collect for that frequency, and statistically it does help, as you would expect, predict the next quarterly release," he said, but he added that whether monthly GDP data shows mildly negative or positive growth of limited use.

 "On average, these days, annual growth in the UK is about one-and-a-quarter percent, quarterly growth about point-three, monthly, about point-one, right? And then the volatility in the monthly date is about point-two. So one third of the time you've got a minus-point-one, and I don't think when you've got a minus-point-one, that's something to write a headline about personally," Benford said.

 "It was for that reason we moved to three-month-on-three-months -  just smooth through some of that noise," he added.