By Sophie Barrand, Ellis Daniel, Sumit Dey-Chowdhury and Andrew Walton, Office for National Statistics
In 2017, the Office for National Statistics (ONS) introduced Value Added Tax (VAT) data as a key data source in our short term estimates of gross domestic product (GDP). The focus so far has been on the incorporation of VAT turnover, but this data source also provides information on the purchases made by individual businesses. Today we publish early research exploring how these VAT data may help explain the changing nature of the relationship between turnover and expenditure and their potential to be used as proxies for output and intermediate consumption in the early pandemic period.
As this is our first step in using the VAT expenditure data for research, we have focused on the raw expenditure variable: it is absent of the adjustments that are necessary to bring it into line with the concepts used in official estimates of GDP. However, this approach provides significant and interesting insights which include:
- Providing initial timely insights into how the relationship between turnover and expenditure has changed during the first half of 2020, and the possible implications for initial estimates of GDP
- Enabling us to better understand how to apply necessary adjustments to VAT expenditure so that it can, in the future, be utilised more fully within the UK National Accounts
- Informing our understanding of the VAT expenditure composition and distribution to apply any future editing rules to clean the dataset
There are numerous challenges in analysing the VAT returns, specifically in bringing these into line with the concepts of the UK National Accounts. Further research is needed to refine these data and differentiate between total and intermediate expenditures. However, this exploratory research published today offers opportunities that might improve our understanding of how we have recorded early estimates of GDP through the COVID-19 pandemic. The use of VAT turnover and single extrapolation in short-term GDP estimation relies on the relationship between turnover, output and intermediate consumption being stable. This assumption is open to challenge in times of significant economic change and our research has shown this assumption may not have held during the pandemic.
Our research looks closely at businesses’ VAT expenditure in comparison to their VAT turnover and we calculate the concept of gross value added (GVA), defined as VAT turnover minus VAT expenditure. Where firms are reporting “negative GVA” this implies a turnover to expenditure ratio of less than 1. Overall, in the period covered by the COVID-19 pandemic, our research shows that there has been a leftwards shift in this ratio compared to the same quarter a year earlier – that is, in Quarter 2 (April to June) 2020 the ratio declined.
The data imply that there has been a larger adverse effect on turnover than expenditure for firms in the first half of 2020. For example, the proportion of firms reporting negative vGVA increased from 15.5% to 19.8% on the year.
How does this shift manifest within each sector? As the coronavirus restrictions tightened, it is expected that this would likely impact people-facing service industries rather than those trading online. Ultimately, shifts have been prevalent in every sector and these are detailed within our full report. But change has been more pronounced in the services sector, with the proportion with negative vGVA increasing from 13.5% in Quarter 2 2019 to 18.9% in Quarter 2 2020. Less pronounced shifts in the VAT have been seen in construction and manufacturing, supporting broader conclusions that those activities reliant on face-to-face forms of interaction between the producer and user have experienced substantial leftwards shifts in their turnover to expenditure ratio.
There is further work to be done, such as isolating the type of spending, as well as identifying and addressing potential outliers within our expenditure variable. As new data become available, extending this analysis to the latter quarters of 2020 and beyond will allow for an enhanced picture of the longer-term impact on businesses and the sustainability of increased ratios of expenditure to turnover. For example, how did our panelled population change their expenditure as the pandemic continued? How is this reflected in later lockdowns? Are smaller businesses or certain industries substantially affected?
The timeliness of the VAT data has facilitated prompt analysis and given earlier insight into these changing relationships that would have been possible with traditional sources. The increased capability and use of administrative data have allowed us to interrogate movements at the firm level while providing an overall view of the whole economy. We are taking steps to address the challenges faced in more fully utilising this important data source and this work paves a way, with the scope to triangulate this work with surveys and other administrative data sources. This will provide valuable insight into each of the sources, how they compare and how they can be optimally used to inform the UK National Accounts.
We value feedback on our initial approach and findings.
You can download and read a copy of the Technical Report in full here.
Sophie Barrand and Ellis Daniel are in the Economic Microdata Research team, Sumit Dey-Chowdhury in Macroeconomics, Statistics and Analysis and Andrew Walton in National Accounts Coordination at the Office for National Statistics.
ESCoE blogs are published to further debate. Any views expressed are solely those of the author(s) and so cannot be taken to represent those of the ESCoE, its partner institutions or the Office for National Statistics
