Impact of trade in goods data collection changes on UK trade statistics: adjustments to 2021 EU imports estimates

Since the UK left the EU, the arrangements for how the UK trades with the EU have changed. We are keen at the Office for National Statistics (ONS), working with colleagues in HM Revenue and Customs (HMRC) and elsewhere, to keep users updated on the impact of changes to trade data collection on our trade statistics and how we plan to account for these impacts.

In our previous articles on the Impact of trade in goods data collection changes on UK trade statistics, we have provided information on the data collection changes, as well as detail of the discontinuity between the two compilation methods.

To account for these discontinuities, we have now applied adjustments to our estimates of goods imports from the EU for the period January to December 2021. These adjustments have been applied to ONS UK trade statistics in line with the National Accounts Revisions Policy, with revised estimates incorporated into GDP quarterly national accounts, UK: July to September 2022 on 13 January 2023.

This article summarises these adjustments to our estimates of goods imports from the EU in 2021. It also details further work that is ongoing to review the pre-2021 discontinuity in trade in goods with the EU as well as the impact of Staged Customs Controls in 2022. We will keep users informed of any further changes as we continue to provide our best estimates of trade.

In January 2022, HM Revenue and Customs (HMRC) implemented a data collection change affecting data on imports from the EU to Great Britain (GB). HMRC investigated the change in data collection methods for GB imports from the EU, concluding that, when comparing the components that form the trade in goods data in 2021 and 2022, there was a discontinuity of around 6% by value between the two compilation methods.

This discontinuity was caused by a combination of data not previously collected under the Intrastat survey. These data were:

• traders with an Economic Operators Registration and Identification (EORI) number, who are not Value Added Tax (VAT) registered and therefore not in the 2021 dataset

• traders with an EORI number, who are VAT registered but not found in the 2021 dataset

• trade from parcel post

More detail on the findings of the investigation can be found in our article, Impact of trade in goods data collection changes on UK trade statistics: 2021 to 2022.

After concluding the investigation into the impact of the data collection change on goods imports from the EU and considering possible options, we have now implemented adjustments to our estimates of goods imports from the EU for the period January to December 2021. These adjustments account for data which were not previously collected under the Intrastat survey.

As the data collection change for imports from the EU to Great Britain (GB) moved from the Intrastat survey to customs declarations in January 2022 – 12 months after the equivalent change for GB exports to the EU – estimates of imports from, and exports to, the EU were on a different basis to each other in 2021. The adjustments made by the Office for National Statistics (ONS), described in this article, bring the two series onto a like-for-like basis such that the impact of the change of data source is now seen at the same point in time for both imports from, and exports to, the EU. Additionally, the revised estimates are now in line with the equivalent statistics as published by EU countries, who also switched to customs declarations for their reporting of UK trade in both directions in 2021.

The adjustments represent an overall uplift of 6.6% from the previous estimates, which equates to 6.2% of the new estimates (which we refer to as uplift contribution to revised 2021 EU imports). This is in line with the previously published discontinuity of approximately 6%.

The adjustment method consisted of two approaches, one applied at chapter level (96 chapters) to account for estimated trade from traders not found in the 2021 dataset. This has been in the form of an uplift percentage applied to all reported trade across the year. There was a separate approach for chapter 99 (other products) to account for the inclusion of parcel post trade into 2021 EU imports figures, an additive adjustment which takes into consideration the monthly variation of parcel post data.

Adjustments have been applied at a chapter level and can be seen in the ONS' published commodity series using the Standard International Trade Classification (SITC) and the classification of product by activity (CPA) commodity classifications. The data published by HM Revenue and Customs (HMRC) (including those published at a chapter level) have not been adjusted.

Adjustments were applied at chapter level for all chapters other than chapter 99 (other products). Multiple data sources were used to calculate lower and upper bounds and mid-point options for the size of the uplift adjustment and followed a decision-tree approach for deciding the most suitable option for each category.

The lower bound consisted of the share of imports in 2022 from traders with an Economic Operators Registration and Identification (EORI) number, who are not Value Added Tax (VAT) registered and not in the 2021 dataset. The Intrastat data collection did not capture any data for non-VAT registered traders.

The upper bound consisted of the lower bound, plus the share of imports in 2022 from traders with an EORI number, who are VAT registered but not found in the 2021 dataset. While considered to be "not found", and while many traders within this VAT-registered grouping were not captured at all in the 2021 data, there is some anecdotal evidence to show that some may have been included in the 2021 data. This may have been either through using a different trader name or EORI number, or the trade may have been reported by a different entity.

We have worked closely with HM Revenue and Customs (HMRC) colleagues to understand the nuances of the data and trader behaviour and decided that the mid-point between the lower and upper bounds would be an appropriate starting point to apply uplift adjustments.

Using the mid-point as the starting point, decisions were made on the scale of the uplift adjustment at chapter level, using additional data sources to validate decisions. The two additional data sources used in this analysis were:

• asymmetry data from Eurostat, comparing total EU level for each chapter

• trade declared through raw customs import declaration data in 2021 (this data source is not otherwise used for contributing to our statistical outputs)

The mid-point of the lower and upper bounds was considered the default decision unless the additional data sources suggest an adjustment of the lower or upper bound values would be more appropriate. As a consequence of this detailed analysis, some chapters were adjusted by the lower or upper bound values, while the remainder had the mid-point uplift adjustment applied (Figure 1). Once the uplift percentages were determined for each chapter, these were applied consistently to all EU imports in Office for National Statistics (ONS) data processing; that is, the same uplift percentage was used for each month of 2021.

Imports of parcel post from the EU have been identified to account for 0.5% of EU imports trade in 2022. Parcel post is the only component of chapter 99 (other products) that we are applying an adjustment to. Our adjustment for parcel post has therefore been to include an adjustment that will lead to it accounting for 0.5% of trade in 2021. Non-EU imports data demonstrate that parcel post is a seasonal series, so this has been used to inform the monthly levels for this adjustment.

After these adjustments have been applied, the net impact is an overall uplift of 6.6% to goods imports from the EU in 2021. The subsection of the trader population with uplift adjustments applied accounts for a 6.2% share of the revised total imports of goods from the EU in 2021 (Figure 2). This is in line with the previously published discontinuity of approximately 6%.

As these adjustments have been made at the chapter level and aggregated to commodity level, the uplifts vary at commodity level (Table 1). Crude materials, and food and live animals, saw the largest uplift adjustments proportionally, with upwards revisions of 12.6% and 11.7%, respectively.

Adjustments have been made at chapter level, producing overall uplifts at commodity level. For example, within the food and live animals commodity (Standard International Trade Classification (SITC) 0), vegetables and fruits saw the largest revisions, of £1,026 million (16.9%), which results in the uplift contributing 14.4% to the revised total (Figure 3).

Vegetables and fruits is composed of products from several chapters, predominantly chapter 20 (preparations of vegetables, fruit, nuts or other parts of plants), which accounts for 35.7% of this commodity, chapter 7 (edible vegetables and certain roots and tubers), which accounts for 35.6% of this commodity, and chapter 8 (edible fruit and nuts; peel of citrus fruits or melons), which accounts for 27.3%. The remaining 1.4% of the vegetables and fruits commodity comes from a small number of products from chapters 11 (products of the milling industry; malt; starches; inulin; wheat gluten), 12 (oil seeds and oleaginous fruits; miscellaneous grains, seeds and fruit; industrial or medical plants; straw and fodder) and 19 (preparations of cereals, flour, starch or milk; pastrycooks' products).

Chapters 7 and 20 received lower-bound uplifts of 18.7% and 15.3%, respectively. The lower-bound option was chosen here, as the other options would have led to a clear divergence from previous trends of asymmetry comparisons with exports to the UK as reported by EU countries. Chapter 8 received a mid-point uplift of 17.6%, as the asymmetry and raw customs declaration data did not provide a strong consensus for choosing an alternative bound.

Our accompanying dataset, providing the uplift percentages by chapter, is available.

Impact of adjustments to parcel post

Our adjustment for parcel post ensured that this accounts for 0.5% of EU imports trade in 2021, which aligns with 2022 proportions, and data on imports of parcel post from non-EU countries have been used to inform the monthly path for this adjustment. Parcel post sits within the unspecified goods commodity (SITC 9), which saw a revision of £1,159 million (113.2%), and the uplift to parcel post imports from the EU, which was not collected in 2021, is the primary driver of the adjustments to unspecified goods (Figure 4). After the adjustments have been implemented, imports of parcel post accounted for 53.1% of the total imports of unspecified goods from the EU.

Unspecified goods also includes chapter 93 (arms and ammunition; parts and accessories thereof) and a small amount of chapter 71 (natural or cultured pearls, precious or semi-precious stones, precious metals, metals clad with precious metal, and articles thereof; imitation jewellery; coin) and 87 (vehicles other than railway or tramway rolling-stock, and parts and accessories thereof). However, uplifts adjustments to these chapters have had a minimal impact on the overall uplift to unspecified goods.

UK imports from the EU pre-2021

The adjustments described in this article apply for 2021 only and bring imports and exports statistics onto a like-for-like basis. The full timeseries for imports from the EU still contains a discontinuity, but now from January 2021 rather than January 2022. We will work with HM Revenue and Customs (HMRC) to consider possible options to account for this discontinuity.

UK exports to the EU

In January 2021, HMRC implemented a data collection change affecting data on exports to the EU from Great Britain (GB). HMRC has investigated the change in data collection methods for UK exports to the EU, concluding that, when comparing the components that form the trade in goods data in 2020 and 2021, there is a discontinuity of around 5% by value between the two compilation methods.

This discontinuity is a combination of data not previously collected under the Intrastat survey, such as parcel post (0.3%) and non-Value Added Tax (VAT) registered traders (1.2%), along with a group of traders using a new or different identifier for their customs declarations (approximately 3.5%).

More detail on the findings of the investigation can be found in in our article, Impact of trade in goods data collection changes on UK trade statistics: 2020 to 2022.

Work is ongoing to review the discontinuity in goods exports to the EU, and we are considering possible options to account for this discontinuity. We will keep users informed of any further changes in data collection.

Staged Customs Controls

In 2021, the use of Staged Customs Controls (SCC) allowed customs declarations to be reported up to 175 days after the date of import for imports of non-controlled goods from the EU to GB. Full customs controls were introduced in January 2022, and as such, July 2022 marked the first full month of data where delayed customs declarations submitted under SCC could not be included. Temporary arrangements still apply for imports of goods from Ireland to GB. While it is not possible to distinguish which customs declarations are associated with SCC, we are reviewing the 2022 import data to determine any effect of SCC. We will keep users informed of any further changes in data collection.