Summary
The decrease in size of standard consumer goods combined with constant prices is a phenomenon real inflation known colloquially as ‘shrinkflation’. This is a topic that many consumers would tacitly recognise as occurring at a rate high enough as to constitute a ‘trend,’ but more exact characterisations are rarely factored into public discussion of the issue. Headlines such as ‘Brexit bites: more than 200 products subject to shrinkflation, says ONS’ may be attempting to encapsulate consumer pessimism connected to the pressures of Brexit on the UK economy, as well as reference a larger global trend among firms seeking cost savings against a backdrop of stable inflation and wage stagnation.
The headline is misleading at best, something acknowledged by the author in the article. The ONS found no link to Brexit in the size changes of products, as displayed by the charts above, and further, the sample size (37 thousand plus) indicates the issue is far from widespread. This underscores the importance of a measured, honest analysis of issues which ‘feel’ true but may not be salient economically. Such sensationalism serves to further erode trust in economists and policy-makers and could be accused of ‘stoking Project Fear’.
What does the chart show?
The bar chart displays price changes among products which underwent size changes between 2015-2017, and details whether these products increased, decreased, or maintained a constant price. The data collected for this chart measures individual price changes, so should one particular product change price multiple times or be recorded at different prices in multiple areas; it would feature more than once in this sample. The pie chart displays the total number of food products which changed size in 2016. Food products are of particular relevance to this analysis as they represent 70% of all CPI consumer goods factored into this data set that changed in size. The blue proportion is the number of products which did not change in size, the orange proportion is the number of products which decreased in size and the grey proportion is the number of products which increased in size.
Why is the chart interesting?
Between September 2015 and June 2017 there were 206 individual products identified as having shrunk. Centering only on this fact, the headlines could be considered accurate. However, to contextualise this, the researchers analysed in excess of 37,400 consumer goods akin to those which would feature in the debate around shrinkflation. The most relevant group of goods in this sample were food products, which constitute around 70% of the size changes. Among a 2016 sample of 37,408 food products, the researchers found a total of 361 which had shrunk in size, only a 1-2% proportion of the total. Although basic, information such as sample size is integral to any conversation about trends such as this and its omission is misleading.
An additional part of the headline and broader narrative around shrinkflation is concern about the impact of Brexit and its perceived effects on consumer goods. According to the sources quoted in the Guardian, sterling depreciation has put supply side pressures on the consumer goods market as imports are now relatively more expensive. From an economic theory perspective, this is certainly an intuitive point to make. However this ignores what the ONS study plainly states in its first paragraph, that ‘there was no trend in the frequency of size changes over this period, which included the EU referendum’. This information is critical to understanding the study and is mentioned later in the article, but the headline disingenuously leads the reader to believe that, per the ONS, Brexit is a, if not the, factor driving shrinkflation.
These inconsistencies highlight a more general need for measured, honest analysis of such issues. These topics are certainly on the mind of the average consumer feeling the pinch but, there is little technical understanding of their scope and relevance. The ONS are unequivocal about this issue as well, offering multiple qualifying statements about their analysis and its limitations, precisely to prevent extrapolation of the kind offered by some journalists.
Analysis such as this using CPI data presents numerous challenges. Multiple agents collect price data across the UK for general products, in order to identify and price the most popular and relevant products. It is therefor not possible to create an exact measure of how many products shrank, stayed consistent or increased in size. The central researchers did not have precise information on the exact product brands or names, and could only access a description sent in by the roaming data collection
agents. This means the only method by which potential duplicates can be eliminated is to compare the descriptions manually, an imperfect and error-prone method. Additionally, the CPI targets popular ‘easy-to-follow products’. Therefor, the size changes in the sample will not necessarily represent the UK consumer market exhaustively. Resultingly, the ONS researchers explicitly state that they ‘do not recommend making … generalisation[s]’, that suggest their snapshot pertains to the market as a whole.
Further to this, only 46% of the 37,408 food products in the sample have a discernable ‘size’ attribute which could potentially be manipulated by producers. Certain products, including whole fruit and vegetables are included in the sample but cannot be included in the number of products which had a size subject to manipulation.
Although food products make up over 70% of the size changes among the consumer goods in the CPI, size changes of personal care products and non-durable household goods were also relevant as there were 49 size decreases and 15 size increases between 2015-2017.
The variation in the figures demonstrates difficulties in analyses such as these. Although the study concluded that from 2015-2017, there were a total of 206 size decreases among all products, it also presented 2016 data suggesting that there were 361 size decreases among just food products. These two figures appear contradictory, further underscoring the need to highlight not only potential duplicates in the sample, but products undergoing one or more size changes during the time period and products whose size changes were recorded at different times/locations by different CPI analysts. Ultimately, the smaller number (206) was reached following the removal of duplicates and redundancies. The chart displaying price changes suffers from similar limitations.
Ultimately, the presence of qualifying statements such as these does not detract from economic analysis but complements it in terms of transparency. Caution must always be exercised when making definitive statements about economic developments, something especially true of unwieldy, nationwide aggregations of trends in markets as large and diverse as consumer goods.
Week 4, 2019