UK retail sales in December were the weakest in a decade, barely rising from a year before. So said the industry’s lobby group, the British Retail Consortium. Statisticians at the Office for National Statistics said they were up 2.7 per cent over the same period.
The discrepancy highlights the difficulty in trying to track retail sales in an industry that remains highly fragmented in some quarters, and where more trade is moving online.
Retail sales are a vital indicator in an economy such as the UK where consumer spending plays an outsized role. Wholesale and retail trade accounts for roughly one-tenth of the economy, more than the entire manufacturing sector.
Yet measuring growth in the industry accurately is far from straightforward. “The problem is that all retail sales series are flawed to some degree,” said Nick Bubb, an independent retail consultant.
As well as the BRC and the ONS, the CBI business lobby group produces figures and so does the accountancy firm BDO. Barclaycard and Visa publish estimates of consumer spending and other indicators include the “footfall” data supplied by Springboard, which measures shopper traffic.
The obvious problem with footfall is it does not capture online sales — around one-fifth of the total, according to the ONS data — and that it cannot tell whether shoppers are spending or just looking. But all of the measures are struggling to adapt to a transforming high street where shoppers are not only constantly changing how much they buy, but also how they buy it.
The BRC, assisted by advisory services company KPMG, asks retailers to enter sales figures each week via an online portal. Food retail figures are fed in by consultancy IGD. The figures are “cleaned”, according to Rachel Lund, head of insight and analytics at the BRC. But there are no revisions to previous months’ figures and no seasonal adjustments.
BRC figures are widely reckoned to be representative of food retailers’ sales; Ms Lund said the data captures about four-fifths of the grocery market, which is highly consolidated and where most sales are still made in shops.
But economists regard the BRC’s numbers for online and non-food sales as less reliable. “It’s mostly just about the big retailers who are members of the BRC,” said Mr Bubb. “It’s fairly widely known that Amazon does not supply its figures,” he added — even though the online giant is now a member of the organisation. The supermarket Tesco, which withdrew from the BRC last year, still contributes food data but not general merchandise numbers.
“The BRC non-food numbers are disproportionately aligned to the likes of Debenhams, House of Fraser and New Look that have underperformed the market. Of the top 10 pure online retailers, they probably have two or three,” said another economist who looks closely at the numbers.
Ms Lund said the BRC captured about two-fifths of the non-food market. “We do receive data from retailers who are not members of the BRC, and we have added new sources in recent years,” she said. However, she added that the BRC could not disclose individual companies owing to confidentiality agreements. Tesco and Amazon also decline to comment.
The ONS monthly survey is mandatory. “We sample all large retailers,” said Rhian Murphy, ONS head of retail sales. Along with a survey of 4,000 small businesses, it means about 5,000 shopping chains are included in the sample, including all the major online retailers. “The BRC sample is 95,” she added.
The official statistics agency “seasonally adjust” its figures, trying to take into account the usual pattern of spending they would see around holidays such as Easter and Christmas. The BRC compares each month with the same month the year before, but this can be difficult because Easter moves. “Even year-on-year comparisons can be quite tricky if you don’t seasonally adjust,” Ms Murphy said.
Partly the difference between the two surveys is because they measure different things. In an article published last year, the ONS said the best comparison to the BRC data is to look at their non-seasonally adjusted, non-inflation adjusted measure that excludes petrol.
Richard Lim, who used to work on the BRC figures before founding consultancy Retail Economics, has attempted to bridge the gap between the two.
“Our model uses different feeds. We take the best parts of each index and adjust for their shortcomings,” he said. “For instance, the BDO series is better than BRC at capturing mid-tier retailers. But the BRC is much more comprehensive for food retail.”
As well as the BDO data, Mr Lim’s model incorporates spending data from payment service provider Visa. He also looks at the market share figures compiled by Kantar Worldpanel, which rely on consumers submitting data about their spending habits.
“We would exclusively look at the ONS,” said Thomas Pugh, UK economist at Capital Economics, a consultancy. The ONS measure is used to calculate overall economic growth, so Mr Pugh looks at the BRC as an indicator of what might happen to the official measure.
Policymakers, like those at the Bank of England, care primarily about economic growth, he said, and so do his clients.
