Skip to content

Identifying underlying patterns in retail sales

How the weather and special events can have an effect on sales data

Retail sales data are published with comparisons based on both the prior month and the prior year. However, the latest publication for August 2013 does not lend itself ideally to either comparison, due to events that affected the data in both July 2013 and August 2012.

The effect of the London 2012 Olympic Games and the heat wave in July 2013

The Olympic Games took place from 27 July 2012 to 12 August 2012, with the football competition commencing on 25 July 2012. The August 2012 period in the Retail Sales Index (RSI) covers the dates 29 July 2012 to 25 August 2012 and hence the vast majority of the Olympic Games. The Paralympic Games, which took place between 29 August 2012 and 9 September 2012, were entirely contained within the September 2012 period in the index.

The Olympic and Paralympic Games were considered a special event by ONS, owing to their significant and wide-ranging impact upon economic statistics. The RSI was flat for the month, but this hides changes within components of the index. For instance, stores selling books, newspapers and stationery reported sharply falling sales, whilst stores selling sporting goods and toys saw strong growth. Notably, the proportion of Internet sales decreased from 9% in July 2012 to 8.3% in August 2012, before resuming their long-term upward trend (Figure 1); feedback from retailers attributed the August decrease to customers watching the Olympics instead of shopping online.

Figure 1: Internet sales as a proportion of all retail sales (excluding automotive fuel)

Figure 1: Internet sales as a proportion of all retail sales (excluding automotive fuel)

Download chart

Meanwhile, July 2013 was the third warmest July on record across the UK. The first three weeks of the month saw the UK's most notable summer heat wave since 2006, with temperatures exceeding 28 degrees Celsius somewhere in the UK every day from the 6th to the 24th. Additionally, there was almost no rainfall throughout England and Wales before the 23rd. Although widespread thundery downpours late in the month meant total rainfall was only slightly below average for July, the total number of rainy days was significantly below average, particularly in south-western areas. As the July 2013 reporting period ended on 27 July, the July 2013 retail sales data did not cover all of these wet days. August 2013 saw more moderate conditions, with slightly above average temperatures and slightly below average rainfall.

How weather affects retail sales

Although the heat wave in July 2013 was not classified as a special event, as the impact on the economy as a whole was not considered significant, retail sales are particularly sensitive to weather conditions. For instance, cold autumn weather will tend to increase sales of winter clothing, while heavy snowfall will reduce sales in most stores due to the impact on transport. Feedback from supermarkets suggested that their increased sales in July 2013 – to the highest levels on record – were driven by sales of alcohol, clothing and outdoor items, as well as food. Two separate reports from business groups – the CBI in late July and the British Retail Consortium in early August – suggested that retail sales growth was driven by footwear and clothing, food & drink and health and beauty products (due to increased sales of hay fever remedies and sun cream). In general, existing summer peaks in demand appear to be exaggerated by heat wave conditions.

Table 1 compares month-on-month seasonally adjusted growth in sales by sector during and after the heat waves of 2006 and 2013 (as the 2006 heat wave started in June, but peaked in July, both months are included). Predominantly food stores (primarily supermarkets) show a consistent pattern of strong sales growth during heat wave conditions, which is reversed when temperatures return to normal. Meanwhile, non-food stores on aggregate saw the opposite pattern to food stores, though to a lesser extent. As the effect was particularly pronounced in food stores in July 2013, the temporary increase in retail sales as a whole was far greater than in 2006.

Table 1: Month-on-month retail sales growth by sector, June-August 2006 and July-August 2013

Chained volume measure, seasonally adjusted

  Percentage change on previous month
  Jun-06 Jul-06 Aug-06 Jul-13 Aug-13
Total retail sales 0.3 0.1 -0.1 1.1 -1
Predominantly food stores 1.4 1.4 -2.1 2.7 -2.7
Predominantly non-food stores -0.2 -0.5 1.1 -0.3 0.4
Non-store retailing 1.8 -6.1 7.8 0.9 -0.2
Automotive fuel -1.9 1.1 -1.1 0.8 -0.6

Download table

Patterns in retail sales data

Given the unusual nature of the month-on-month and year-on-year comparisons, it is necessary to consider broader patterns in retail sales to place the August 2013 data into an appropriate context. Figure 2 illustrates seasonally adjusted retail sales since 2004 in terms of both value (at current prices) and volume (using the chained volume measure). A long period of generally steady retail sales growth was broken in 2008, during the economic downturn that followed the financial crisis; sales in that year were largely flat in volume terms and by the end of the year were falling in value terms. With the exception of a sharp dip in January 2010 due to heavy snow, sales increased in value terms from mid-2009, but remained broadly flat in volume terms, as prices increased more rapidly than earnings with a resultant squeeze on disposable income. Sales volumes grew significantly during the first half of 2013, however, despite this squeeze continuing. By both volume and value, seasonally adjusted retail sales in August 2013 were at the second highest level on record after July 2013.

Figure 2: Seasonally adjusted sales volumes and values, all retailing

Figure 2: Seasonally adjusted sales volumes and values, all retailing

Download chart

Since 2010, annual inflation at predominantly food stores has remained above 3% for most of the period since 2010, whereas annual inflation at non-food stores peaked at 2.4% in the same period and was negative for twelve consecutive months between June 2012 and May 2013 (Figure 3). This is likely to reflect more extensive discounting at non-food stores in response to greater online competition, and significant price rises for staple crops such as wheat – variously attributed to factors including poor harvests and increasing demand for crops as biofuels – affecting food prices. As a result, predominantly food stores saw much faster growth at current prices despite volumes falling slightly between 2010 and early 2013. This is illustrated in Figure 4, which also illustrates how the heat wave effect in the July data was attributable to food stores. There are also several other anomalous one-month spikes in the data at this level. 

Figure 3: Annual inflation, predominantly food stores and predominantly non-food stores

Figure 3: Annual inflation, predominantly food stores and predominantly non-food stores

Download chart

Figure 4: Seasonally adjusted sales volumes and values, predominantly food stores and predominantly non-food stores

Figure 4: Seasonally adjusted sales volumes and values, predominantly food stores and predominantly non-food stores

Download chart

In summary, the August 2013 data, despite their headline figure of a significant month-on-month contraction, do not represent a breaking of an emerging pattern of retail sales growth in volume as well as value during the first half of 2013, a pattern which in turn was exaggerated by strong supermarket sales during the July 2013 heat wave.

Categories: Economy, National Accounts, National Income, Expenditure and Output
Content from the Office for National Statistics.
© Crown Copyright applies unless otherwise stated.