ONS data, trade in non-monetary gold & computing the growth rates of EU & WTO trade: 1999 -

The UK's Office for National Statistics provides by far the best data with which to analyse long-term trends in UK trade. It's current release (accessed March, 2019, on which this site relies) tracks exports and imports with EU and non-EU trade partners back to 1998 for goods, and 1999 for services. By examining data on trade with European Free Trade Association (EFTA) and World Trade Organisation (WTO) partners, by separating out trade with EU Accession States since 2004, and by extracting trade in non-monetary gold, analysts can assess the validity of two broad assumptions:

  • First, that ONS trade data for UK's EU trade accurately reflects UK trade inside the EU Customs Union during the period 1999 - 2018

  • Second, that ONS trade data for UK's non-EU trade accurately reflects UK trade with partner countries under WTO terms during the period 1999 - 2018.

All data & calculations, with links to ONS sources, are available here.


This article is designed to help analysts use ONS trade data to investigate long-term trends in UK trade. The basic delineation in ONS trade data - between UK's EU and non-EU trade - neatly divides UK trade into two roughly equal parts.

This site calculates compound annual growth rates (CAGRs) for multiple sectors of UK trade based on that basic EU/Non-EU delineation. The data is deflated using ONS' import and export deflators, then three-year averages attained at each end of the time series, and the CAGRs calculated. For UK's top 10 manufacturing exports, comprising 80.1% of all UK manufacturing exports in 2018, the CAGRs for UK exports are as follows:

Source ONS, created with Michael Burrage. Available on request.

Drawing conclusions from the raw data is subject to three caveats, however.

1) UK's non-EU trade data reflects a variety of different trade relationships. In all categories except 'Chemicals' and 'Food', UK exports more - and usually substantially more - to non-EU markets than to EU. In 2015, according to calculations made by Mr Burrage in 'It's Quite OK to Walk Away' (p.86 - 88) approximately 73% of this non-EU trade was conducted on WTO terms. This portion has changed over the two-decade period for which data is examined, however. For example, when the EU-South Korea trade treaty was provisionally applied in 2011, that had the effect of reducing the portion of non-EU trade conducted on WTO terms. Similarly, the ultra-fast growth of exports to China has the opposite effect. Nevertheless, over one-quarter of UK's non-EU trade is conducted under a variety of arrangements, from EFTA to the partial EU-Turkey Customs Union, to the Lome Convention system of preferences. The CAGRs for these separate portions of trade may be different: indeed, one would expect them to be different precisely because the non-WTO portion should reflect the preferential trade terms negotiated by EU on UK's behalf.

2) The accession of new EU states impacts the growth rates of UK exports. At the start of the time period (1999) the UK was in a Customs Union with just 15 EU states. Ten new states acceded to the EU on 1 May 2004 and joined the Customs Union on the same date: Cyprus, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Slovakia, and Slovenia. Bulgaria and Roumania joined in 2007, and Croatia in 2014. Although taking just 7.4% of UK exports in 2018, the fact that these states joined the Customs Union part way through the period impacts the overall compound annual growth rates (CAGRs) for UK's EU trade. In its trade statistics, the ONS counts these states as belonging to the EU from before they acceded. To be statistically fair, however, analysts need to recognise that the CAGRs for UK's EU trade may be unfairly distorted by including trade with countries from periods anterior to their joining the EU.

3) Trade in non-monetary gold. When UK sells non-monetary gold, the ONS accounts for the transaction as an export of precious metals (which is part of the its Basic Metals category, according to conversations with ONS) and this counts towards overall manufacturing trade statistics. It may be asserted that UK does not accrue the same value from exporting gold as from other manufactured goods, and trade in gold is, by convention, sterilised from calculations of GDP. Therefore, ideally, the trade in non-monetary gold could or should be extracted from trade analysis, to give a more fair or nuanced view of UK trade.

1. UK's non-EU trade and WTO terms

The table below presents the CAGRs for UK exports to its 13 largest non-EU export partners. Their value to UK as an export market in 2018 can be seen in the penultimate column, with declining percentage shares. In total, these 13 trade partners took 78.3% of UK's non-EU exports in 2018, so this sample encompasses more than three-quarters of UK's whole non-EU exports, by value.

The purpose of the table is to analyse whether UK's export CAGRs with non-EU states generally matches the CAGRs of UK trade with countries for which WTO rules apply, as well as UK's principal EFTA partners. In short, it does. This means that CAGRs for non-EU trade are - at the top level - a fair representation of the performance of UK's trade that is conducted on WTO terms, as well as trade with EFTA countries. This table does not allow assessment of trade growth with Lome Convention group countries, or FTA partners as a group - although by implication, it is likely that growth rates are slightly below the overall non-EU average of 3.17% p.a. for the period 1999 to 2018.

The fact that UK's WTO-governed exports appear to have grown very marginally faster than those to countries with whom the UK has EU-negotiated trade arrangements is worth noting - if only in passing. Michael Burrage investigated the performance of UK's EU-negotiated trade agreements in 'It's Quite OK..' this analysis merely confirms that he would appear to be correct, based on entirely different data sources (Michael Burrage based his research on global trade databases).

Source: ONS, supplied on request by Head of Trade Statistics. Calculations: UK Trade_Goods & Services, Tab 6.

Back to the table. Together the value of UK's top 13 trade partners equates to 78.3% of all UK's non EU exports, which were worth £178.4 bn in 2018. Thus it covers over three-quarters of the relevant trade. UK's trade with these countries non-EU countries can be divided into four principal categories:

  • World Trade Organisation (WTO): trade conducted with countries under WTO rules. During this period 10 countries out of the 13 traded with UK predominantly under WTO rules.

  • European Free Trade Agreement (EFTA). trade conducted under the EU's agreements with the members of the European Free Trade Agreement: Switzerland and Norway, make the cut into this table, Lichtenstein and Iceland do not.

  • Free Trade Agreements: trade conducted under EU-negotiated trade agreements. The only such partner to make this top-13 list is South Korea, which agreed a EU-South Korea Free Trade Agreements in 2011, from which date it was provisionally applied. Additionally, the UK trades with scores of developing countries under the Lome Convention.

  • Hybrid Customs Arrangements: for example, the UK trades with Turkey under the Ankara Agreement, which creates a hybrid Customs Union in non-agricultural products. Turkey also makes this top-13 list.

The table shows that just 21.7% of UK's non-EU exports went to countries not included in this table. The majority are countries that fall under the protections of the Lome Convention, though they also include Mexico and South Africa, with whom the EU now has free trade agreements (FTAs).

To isolate the CAGR for UK's WTO trade, the countries are divided between those with whom the UK has mostly conducted trade under WTO terms since 1999 (there are 10), and those with whom the UK has not: Switzerland (EFTA), Turkey (hybrid customs union) and Norway (EFTA). South Korea is included in the WTO category, since for 3/5ths of the time series, UK trade was conducted on WTO terms, though account has been made for the eight years when it was not. Altogether, and using an highly simplistic, pro-rata calculation, it is reasonable to say that 97% of the trade growth categorised here as WTO, can be attributed to trade that was conducted on WTO terms during 1999 to 2018.

The average CAGRs for this group is 3.21% p.a. (pink), which is 0.04% p.a. faster than the 3.17% p.a. (green) overall CAGR for UK's non-EU exports. Together these top-ten trade partners contributed 69.3% of UK's non-EU exports in 2018. Using calculations provided by Mr Burrage (Its Quite OK to Walk Away, pp 86-88), this total of 69.3% of UK's non-EU trade is within 4-5% of the estimated total of 73% of UK's non-EU exports that were conducted under WTO rules in 2015.

Given that less than 10% of UK's WTO trade (more likely 7%) is missing from the calculations, this makes the figure highly compelling as a metric of the CAGR of UK's WTO trade - and whatever the CAGR of UK's WTO partners not included in the table, it reveals the success of UK trade under WTO rules for its ten, or perhaps nine largest WTO partners.

Again, it should be noted that exports to Switzerland, Norway and Turkey (blue) have grown very slightly slower than with UK's principal WTO trade partners at 3.19 %, though again, this is till fractionally faster than the overall non-EU export CAGR. The 21.7% of UK's non-EU exports not defined in this table grew at a below-overall rate, at 3.1%.

Thus so far as this table is able to calculate, and on the basis of an examination of 78.3% of UK's non-EU trade, it is safe to make the following assertions:

  • that UK's goods exports traded under WTO rules appear to have grown very marginally faster during 1999-2018 than UK's goods exports to the principal EFTA countries plus Turkey

  • that UK's goods exports traded under WTO rules appear to have grown faster during 1999-2018 that the average for UK's non-EU exports

  • consequently, that it is safe to equate the CAGR of UK's non-EU goods exports - as presented by the ONS - with the CAGR of UK exports to countries under WTO rules from 1999-2018.

2. The complicating factor of EU accession states

As published, the ONS' EU trade data does not technically represent trade that UK has conducted with the EU since 1999, even though it is classified as such. Ten new states acceded in 2004, two in 2007 and one in 2014. The ONS has chosen NOT to extract these states from the times series data but present the data as if the were members from 1998. After consideration, the author concurrs for the following reasons.

1. Scale. Even after the accession of 13 new members, the share of UK's exports which they took in 2018 was just 7.4% (see diagram below - exports to accession states only kick in at the pink wedge, valued at 6%). The EU 15 - the members of the EU with whom traded via the Customs Union for the entire 1999-2018 period of this sites analysis - still account for 92.6% of UK's goods exports. Thus, the impact which the inclusion or exclusion of the accession states can have on the overall CAGR for UK exports to EU is marginal.

Source: ONS. Calculations: UK Trade_Goods & Services , Tab 5

2. Trajectory of growth. Scrutiny of the trade data for each accession country before and after accession shows no dramatic change. Exports were growing pre-2004, and they carried on growing post 2004, though exports to Poland did pick up pace post accession and Poland is easily UK's biggest export market among the accession states. There is however, no definitive change in the trajectory of UK's exports to accession states that is observable from the year of their accession. UK exports to the 2004 accession states grew over the five-year period prior to their accession (by £1.2 bn) and continued to grow (by £6.6 bn) in the fifteen years following. It may be deducible that exports accelerated post accession, but the author's methodology cannot be applied to a period as short as five years.

The essential point, however, is that the accessions states were already exports markets prior to EU accession (purchasing approximately £4.5bn of UK goods), and that exports were already growing, so that it would be misleading to simply add the values of of those exports to UK's EU export totals, from the dates of accession, and claim they represented the experience of UK's trade with the Customs Union.

3. The accession effect

Nevertheless, even if the CAGRs were constructed only for the period when the accession states were members, the question of logic arises. Are they fairly included when the question that is being posed of the Data is: ‘Has the Customs Union benefited UK exports.’

One could argue that since most of the accession states have been members for most of the time series, then inclusion truly reflects the evolving nature of the Customs Union. Equally, one could argue that what the data actually captures is an ‘Accession Effect’, i.e., the one-time consequence of liberalising trade. And in fact, there was a definite spike in UK goods exports to the 2004 accession states in 2006.

Source: ONS. Calculations: UK Trade_Goods & Services , Tab 5

If there was a one-off 'Accession Effect' then inclusion distorts the CAGRs, because it captures an unrepeatable event, rather than the ongoing value of tariff free trade to UK – which is the questions being asked of the data.

A safe expedient would be to extract all accession states from the data series,and calculate CAGRs only from the 92.6% of UK exports that actually traded within the Customs Union during the 1999-2018 period. The CAGR of UK's goods exports to those EU 15, however, drops briskly to just 0.08% p.a., from the overall rate of 0.31%. (see below).

Source: ONS, Calculations in UK Trade_Goods & Services, Tab 5

Consequently, the author is content to follow the ONS’ example, and count the accession states as if they were members of the EU - and by default the Customs UNion - during the entire period. In doing so, it is estimated that inclusion may very fractionally raise or lower the actual CAGR of UK goods exports into the Customs Union during the period 1999 to 2018. Given that the EU accession states have prove good customers for UK exporters, the author prefers to include them in all calculations, and allow UK-EU trade data the benefit of statistical doubt.

The only statistical alternative open to the author is to omit them entirely, which further reduces the CAGR of UK exports to the EU.

3. Trade in non-monetary gold

Some analysts correctly point out that trade in non-monetary gold distorts export statistics. The point is a fair one: unless the UK is mining it, the export of gold resembles the transacting of specie for cash, rather than the sale of a created good.

In ONS data, however, trade in non-monetary gold is accounted for within the 'Basic Metals' category of manufactured goods. Its line item is 24.4 'Basic precious and other non-ferrous metals'. Consequently, this category captures the effect of UK trade in gold, and so the flows of UK non-monetary gold find their way into UK's manufacturing data and goods data.

The question then arises: to what extent does the inclusion of trade in gold impact the trade data?

Since ONS cannot supply trade in gold as its own line item, the safest method to assess its general impact is simply to extract the entire 'basic precious and other non-ferrous metals' line item from UK's overall goods trade data and then observe the change. The author notes, however, that this approach, is highly simplistic, in that the category includes metals other than gold. It is, however, the only method of 'sterilising' the ONS data so that the entire effect of trading in non-monetary gold is extracted.

Adopting this statistically scorched earth approach reduces UK's non-EU exports by about £7-8 billion per year in each of the last three years, and UK's EU exports by £ 3-4 billion (though imports of both also grew). Exports of precious metals have grown more quickly outside the EU during the 20-year period, however, in particular, the author understands, to Switzerland, China and India.

Source: ONS, 2019. Calculations in UK Trade_Goods & Services, Tab 7

Nevertheless, extracting it entirely from UK's good export data for UK 1999-2018 does reduce the UK’s goods export CAGR to non-EU countries by 0.15 %, from 3.17% p.a. to 3.02%. Extracting the impact of trade in non-monetary gold on UK’s EU exports results in a smaller reduction – of 0.04%, or from 0.31% p.a. to 0.27% p.a. Thus, the overall effect is slight. And since most of this site is dedicated to sectoral analysis, nine-out of the ten sectors analysed are unimpacted in any case.

The author notes in passing, that if one were to combine consideration of the effect of trade in non-monetary gold and the accession states, then the CAGR for UK exports to the EU 15 would drop almost literally to zero - or 0.04% p.a. - for the years 1999 to 2018.

The case for further study

Unfortunately, the ONS cannot supply details on trade in precious metals by country. Thus, it is impossible to answer basic questions relating to its specific impact on trade, besides registering its nominal impact on CAGRs. It is impossible, for example, to assess the impact of CAGR calculations for WTO partners, as opposed to members of EFTA. (If, say, half of all UK's non-EU exports of precious metals was exported to China in 2018, that would equate to approximately 25% of UK goods exports to China. Nevertheless, the CAGR of UK exports to China would still be 10.5% p.a. for the time period, as opposed to 12.3% p.a.)

On a different tack, if the export of non-monetary gold is to be considered a negative for UK trade - which is a questionable assumption - the UK can at least take comfort from its non-EU trade. Unlike in its EU trade, imports exceeded exports by £1.76bn in 2018. Which means, however irrelevant it is to manufacturing, UK currently has a surplus on its trade in gold – at least with countries outside the EU. Inside the EU, the UK sold £1.24 bn-worth of gold more than it bought.

[6th June 2020]

In February 2020 – and presumably in response to queries about the impact of trade in non-monetary gold – ONS helpfully published more precise data, using the Standard International Trade Classification (SITC) categorisation. This confines the ‘Precious Metals’ classification to just gold, silver, platinum and palladium, while still measuring it on a BoP basis. These values are far lower - just £1.655 billion for all exports in 2018. This means the potential impact of trade in non-monetary gold as measured on a BoP basis is effectively capped, in so far as the CPA data used here is concerned. But this publication provides analysts with no indication of who the precious metals were being sold to. This is vital, since this analysis relies on computing per-country export CAGRs, which are then rolled into trade-partner type aggregates.

Fortunately, another ONS trade dataset 'Country by commodities[i]' does provide per country data. This dataset conforms to different categorisations but is still constructed on a BoP basis. For confidentiality reasons, ONS is obliged to categorise gold in its full commodities publication as 'Unspecified goods', but it does provide ‘per country’ values. Whilst again it is not possible to determine the proportion of gold within this category, rocketing trade in non-monetary gold in 2019 provides the necessary clue. In 2019, exports of 'Precious Metals' in SITC classifications shot up from £1.655 bn in 2018, to £14.954 while exports of unspecified goods in commodities listings rose from £2.98 bn to £16.4 bn. This illuminates the composition of the ‘Unspecified Goods’ category and so unlocks the challenge of ascribing values to specific countries.

An update to this article providing valuations of exports of non-monetary gold to principal UK trade partners will be forthcoming. Nevertheless, it should be noted that according to this new ONS data, exports of non-monetary gold were below £4 billion for the 2016-2018 period, far less than the £7-8 billion estimation used above, which would have included trade in processed urianium.

4. ... and finally: the big picture

Before closing, it is worth remembering why CAGR calculations are important. The basic question the author wishes to ask of ONS data is: "Has membership of the Customs Union benefited UK trade." Given that the UK's non-EU trade data accurately reflects UK's trade on WTO terms, and that the accession of new states and trade in non-monetary gold only fractionally impact the percentages, it would appear not. Since 1999, UK's EU exports have grown by approximately 0.31% p.a., and UK's non-EU exports have grown at 3.2% p.a.

But what if this simply reflects an economic reality - that the EU economy is growing far more slowly than the UK's non-EU global markets?

In 'It's Quite OK to Walk Away' Michael Burrage analysed multiple international trade databases and found that this could not possibly explain UK's slow export growth rates with the EU. Virtually every developed economy, he concluded, had managed to grow goods exports to EU faster than UK, many of them by trading under WTO rules.

The author's own analysis agrees with this conclusion. On the basis of official US trade data, it appears that the United States has - while trading on WTO terms - increased its goods exports to the EU by 2.24% p.a. over the period 1998 - 2017 (See Tab 4 in UK Trade_Goods & Services). For all the vast political energy UK has expended on its EU membership since 1999 - plus the financial cost, however calculated - the net result was to get comprehensively overtaken by the US as a goods exporter into EU markets.

Source, ONS & US Census: UK Trade_Goods & Services_Tab 4

So a basic question presents itself: if the reason for UK's failure to grow exports to EU isn't the EU's own poor economic growth (see graph above), then what is it? From sectoral analysis - of trade in motor vehicles, aerospace, pharmaceuticals, food and others - it appears that the Customs Union simply provides no commercial advantage to UK in sectors where UK is globally competitive, but provides EU manufacturing with definite advantages in sectors where UK isn't.

The long term impact of that dynamic is already apparent in UK's stagnant goods exports, and crisply accelerating £101.5 bn deficit in manufactured goods. Common sense says this trend will continue unless UK changes the way it trades with the EU.