Do Sales Matter? Evidence from UK Food Retailing

This paper assesses the role of sales as a feature of price dynamics using scanner data. The study analyses an extensive, high frequency panel of supermarket prices consisting of over 230,000 weekly price observations on around 500 products in 15 categories of food stocked by the UK’s seven largest retail chains. In all, 1,700 weekly time series are available at the barcode-specific level including branded and own-label products. The data allows the frequency, magnitude and duration of sales to be analysed in greater detail than has hitherto been possible with UK data. The main results are: (i) sales are a key feature of aggregate price variation with around 40 per cent of price variation being accounted for by sales once price differences for each Unique Product Code (UPC) level across the major retailers are accounted for; (ii) there is considerable heterogeneity in the use of sales across retailers; (iii) much of the price variation that is observed in the UK food retailing sector is accounted for by price differences between retailers; (iv) only a small proportion of price variation that is observed in UK food retailing is common across the major retailers suggesting that cost shocks originating at the manufacturing level is not one of the main sources of price variation in the UK; (v) own-label products also exhibit considerable sales behaviour though this is less important than sales for branded goods; and (vi) there is some evidence of coordination in the timing of sales across retailers insofar as the probability of a sale at the UPC level at a given retailer increases if the product is also on sale at another retailer.