Dear Professor Sir Ian,

Ahead of the next session of International Trade questions in the House of Commons, which has been confirmed for 10th June, I am writing to seek your advice regarding statements by government ministers relating to employment in export-intensive industries at constituency and regional level, which I am concerned may be giving a misleading picture to the public.

In last month’s session, a number of MPs asked questions about the benefit of exports for employment in their local constituencies, and received responses from Secretary of State Liz Truss, and Minister of State, Graham Stuart, which included the following assertions:

“There are over 4,000 jobs in export-related industries in Redcar”; “There are already 2,000 people in Aberconwy working in export-intensive industries”; “[There are] 300,000 export-linked jobs in the north-east.”

These figures were extracted from the Board of Trade’s recent report ‘Global Britain, Local Jobs’ and the adjoining report from the Department for International Trade (DIT), ‘Local Jobs, Trade and Investment’, which uses ONS and other data sources on sectoral employment, regional employment and export intensity.

As you may be aware, the figures quoted in these reports are accompanied by a large number of caveats and qualifications, some of which I have highlighted below:

• The constituency-level figures rely on a combination of data from 2016 and 2019, so neither take into account recent national events, including Brexit and the Covid-19 pandemic, or indeed significant local events, such as the closure of individual plants;
• ‘Section 6: the role of local data’ of the ‘Local Jobs, Trade and Investment’ report says: “[the data] does not tell how many jobs at the local level are…export intensive, because the local sectoral trade intensity can vary significantly from the national average”;
• • Annex 1 of the report repeats the concerns about extrapolating local figures from the national data, and adds a number of further caveats, e.g. “The definition of ‘intensive’ is arbitrary and can result in misleading results”; “The dataset has significant omissions”; “The quality of these estimates deteriorates as the geographies get smaller”; and
• Annex 4 of the report discusses the many caveats associated with the region-level data created by the Fraser Allender Institute of the University of Strathclyde, as discussed by the FAI with ONS officials, including the age of the data they are based on (2014-16), and the inability to separate exporting firms from non-exporting firms within an industry.

Throughout this report, and the accompanying FAI analysis, the authors who have produced these constituency and region-level estimates are clearly at pains to stress the experimental nature of their methods and modelling; the age of, and significant gaps in, their underlying data sources; and the risks involved with extrapolating local estimates from national data.

That is why, for example, the report says that the figures it provides should be used as a “proof of concept” to “demonstrate the potential of the modelling”; that they should act “as a placeholder until more granular estimates become available”; and that better underlying data is necessary to provide these “better quality estimates”.

Given all that, can I ask whether you consider the way DIT Ministers used these figures in the House of Commons last month, and the precedent that sets for how they and their colleagues may use them next month, to be sufficiently reflective of the caveats and qualifications that the authors of the report have attached to their work?

As you will be very aware, if such caveats are not stated up front, or at all, it does not take a big leap for individuals to present the figures in headline form in ways which are seriously misleading. Here are just a few examples of how Conservative MPs have recently used these figures in local newspapers or their own websites and social media channels:

• “New report reveals several thousand Stoke jobs supported by exports: 10.9% of jobs in the constituency are supported by exports.” (Jo Gideon MP, Stoke-on-Trent Central)
• “MP Julie Marson welcomes report showing almost 14,000 jobs in her Hertford and Stortford constituency are supported by exports” (Bishop’s Stortford Independent)
• “New research shows that around 7,600 jobs in Faversham and Mid Kent are supported by exports.” (Helen Whately MP, Faversham and Mid Kent)
• “Exports support 8,885 jobs in Burnley and Padiham, Anthony Higginbotham MP”

It is of course not the job of the ONS to police how individual MPs are presenting figures at local level, nor would it be possible for you to do so, but I hope you will agree that the example set by government ministers speaking at the despatch box is a vital one in persuading their backbench colleagues to use these figures with care.

To that end, can I ask you to consider whether it would be sensible and helpful for your office to ask DIT Ministers to preface their public use of these figures in future with suitable caveats and warnings as to e.g. their age, their incompleteness, and their experimental nature, and – if you agree – to do so prior to the next session of International Trade questions on 10th June.

I believe that will help to send the message that our public debate about international trade and exports should be based as far as possible on accurate facts and statistics, and it would also show due respect for the work done by the DIT’s statisticians and the FAI’s academics, and the care they took in their reports to discuss the limitations of their research.

I look forward to hearing from you.

Yours sincerely,
The Rt Hon Emily Thornberry MP

Related links:

Response from Sir David Norgrove to Emily Thornberry MP – Estimates of export intensive jobs