Economy, Energy and Fair Work Committee
This report sets out the provisions, along with the background and policy context, for a proposed Committee Bill (under Rule 9.15 of Standing Orders) to address the matter of Pre-release Access (PRA) for economic statistics.
Limitations and Strengths of Scottish Economic Statistics, research commissioned by SPICe on behalf of the Economy, Jobs and Fair Work Committee (now the Economy, Energy and Fair Work Committee) and undertaken by 4-consulting, was published in March 2017.
The Committee issued a call for views in June 2017 and appointed Professsor Graeme Roy – Head of Economics at the University of Strathclyde and Director of the Fraser of Allander Institute – as adviser to assist on its economic data inquiry.
In How to make data count: improving the quality and coverage of our economic statistics, a report published on 12 February 2018, the Committee made 29 recommendations. Most were for the Scottish Government, others directed at the Scottish Fiscal Commission (SFC), Office for National Statistics (ONS) and HMRC.
The Scottish Government’s response to the report was received on 18 April 2018 and accepted the main thrust of the recommendations. However, there were several issues to follow up, the most intractable being PRA.
The prevailing view and recommendation of the Committee was that the practice should cease for four specific areas of Scottish economic statistics: Scottish GDP, the Retail Sales Index for Scotland (RSIS), Quarterly National Accounts Scotland (QNAS) and Government Expenditure and Revenues (GERS). The preference of a minority of members was that there should at least be a presumption against PRA. (See paragraphs 46 and 47 for full text of the recommendation and minority view.)
The Scottish Government did not accept the recommendation or the minority view and there followed extended correspondence over a period of six months, the last of which we received on 21 November 2018.
The Committee led a debate in the Chamber on 22 November 2018, addressing the twin topics of data and economic performance, during which it became evident that the Scottish Government had not altered its view on PRA.
Having considered its next steps at a meeting on 29 January 2019 and again on 12 March 2019, the latter with input from the Parliament’s Non-Government Bills Unit (NGBU), the Committee decided that an informal working group should look at the policy options for a possible Committee Bill and feedback its findings to the wider committee.
That group – made up of the Convener, Deputy Convener, Jackie Baillie and Andy Wightman, with support from the Clerking team and SPICe – met twice to consider the approach it wished to recommend to the rest of the Committee.
At its meeting on 14 May 2019, the Committee discussed its approach and approved the content of this proposal report for a Committee Bill.i
At that meeting, the Committee decided to delay publication of its proposal report whilst seeking further views from the Cabinet Secretary for Finance, Economy and Fair Work. The correspondence can be found here.ii
PRA is the practice of making official statistics, and the written commentary that accompanies them, available in advance of publication to specific individuals not involved in their production. It is granted by the “person responsible” (as defined by section 67 of the Statistics and Registration Service Act 2007); for instance, the ONS’s National Statistician or, in the case of official statistics produced by a UK Government department, the minister in charge of that department. The person responsible may grant pre-release to an eligible person. In most cases this includes government ministers and officials who advise them.
The Pre-Release Access to Official Statistics Order 2008 covers all statistics which are not wholly devolved and therefore all UK Government departments and any other public bodies in the UK that produce non-devolved statistics, including ONS. The “eligible people” who may benefit from pre-release under this order include UK Government as well as Scottish Government ministers. The UK Order grants PRA for a maximum of 24 hours prior to publication. In exceptional circumstances, PRA may be granted in excess of 24 hours if the person responsible believes the likely benefit to the public would outweigh the detriment to public trust in official statistics likely to result from such an access.
PRA in Scotland comes under the Pre-Release Access to Official Statistics (Scotland) Order 2008. This Order is made under the authority of section 11 of the Statistics and Registration Service Act 2007 and covers wholly Scottish devolved official statistics. It allows PRA—
For market-sensitive statistics, a maximum of one working day before publication. “Market-sensitive” is defined in the Scottish Order as statistics which “when disclosed would, in the opinion of the person responsible, be reasonably likely to have a significant effect on the value or traded volume of any investment”, and
For other official statistics, a maximum period of five working days before publication.
This Order also states that PRA may be increased by the person responsible, Scotland’s Chief Statistician, to an unlimited extent if they deem it necessary to fulfil the reasons justifying the granting of PRA as set out in the Order e.g. ensuring that ministers are able to comment publicly on the statistics. In doing so, they must take into account the risk of detriment to public trust in the integrity of official statistics. The Order requires that the person responsible must publish the reasons behind their decision as soon as reasonably practicable.
The justification for PRA has typically been that it is preferable for Ministers to be carefully briefed on the statistics in advance of release so that they can make sensible and informed comments at the time the figures are published. In addition, many senior statisticians in government spend only a relatively small part of their time actually working on the production and publication of statistics. They often have a separate responsibility to offer confidential advice to Ministers and officials on the implications of the statistics for policy and decision-making.
Others argue that equality of access to official statistics is a fundamental principle of statistical good practice, and the existence of pre-release access leaves ministers in a privileged position and undermines trust in our official statistics system. While there is no evidence of the statistics being subject to “influence” prior to their publication, the greater concern is around the opportunity to “spin” the numbers in a positive way in advance of any other commentator or political party being able to respond.
In March 2010, the UK Statistics Authority (UKSA) published Pre-Release Access to Official Statistics: A review of the statutory arrangements, which included a discussion of the arguments for and against.
The review was critical of the “widespread expectation” that Ministers should comment immediately when statistics are published and commended a “central principle of good statistical practice – equality of access”. It stated—
All the parties to public debate should ideally have the same statistical information available to them at the same time and only the minimum number of people should see statistical results before they are published.
Pre-Release Access to Official Statistics: A review of the statutory arrangements, March 2010
It also sought a shorter maximum period of pre-release access than the 24 hours at the UK level and “much shorter than the 5 days allowed in Scotland and Wales”. Furthermore—
We believe it would be in the public interest if all UK administrations amended their secondary legislation to adopt a maximum period of pre-release access of 3 hours, with a shorter period as the norm. A three-hour limit was also recommended by the House of Commons Treasury Committee in 2006.
Pre-Release Access to Official Statistics: A review of the statutory arrangements, March 2010
It contended—
Three hours is time enough for Ministers and officials to understand what is to be published but short enough to ensure little opportunity to influence or exploit the information being released. It may be necessary to introduce such a change in stages, perhaps starting with those statistics for which strict control of pre-release access is most important and practical. This would include key economic statistics and the results of major surveys.
Pre-Release Access to Official Statistics: A review of the statutory arrangements, March 2010
On 1 July 2017, ONS ended all 24-hour PRA for its official statistics. The National Statistician wrote to Chair of the UK Statistics Authority Board setting out the decision—
On the basis of all the information now available to me I consider that the public benefit likely to result from pre-release access to ONS statistics is outweighed by the detriment to public trust in those statistics likely to result from such access.
Letter from John Pullinger to Sir David Norgrove 15 June 2017
It was a development welcomed by the Bank of England, which followed suit, and statistical organisations such as the Royal Statistical Society (RSS). Commenting at the time, RSS’s Executive Director stated—
We hope it will set an example for other government departments to follow and bring forward the day when pre-release access to all official statistics is ended.
Royal Statistical Society press release, 19 June 2017.
The argument that PRA is required so ministers could be “informed rather than ill-informed” was understood by Professor Sir Charles Bean (author of the 2016 Independent review of UK economic statistics). He had been “on the other side of the fence” when a member of the Monetary Policy Committee but said—
Nevertheless, I think that the balance of arguments supports pretty strict limitations on pre-release access.
Economy, Jobs and Fair Work Committee, 7 November 2017
When asked by the Committee during the data inquiry whether PRA should be removed from key economic statistics, Professor Bean, along with UKSA and ONS, agreed it should.
The Scottish Government pointed out during their evidence that the ONS approach is unusual in that no other UK Government departments have followed its example – although it is noted that the Cabinet Office recently undertook a review of PRA for Whitehall departments, the outcome of which was said to be inconclusive.
In evidence to the House of Commons Public Administration and Constitutional Affairs Committee on 22 November 2018, the Chair of UKSA gave his view on balancing the arguments on PRA. John Pullinger said that as President of RSS he had “always argued that fairness demands that everyone has equal access to statistics”. His UKSA role, however, required that he “balance that argument, which is clear, against arguments that Ministers would wish to make statements at the time a statistic is released”. He explained to the Westminster Committee—
…I have carried out quite a substantial review that resulted in a reduction in release but having completed that review, looked at the outcome, and weighed the arguments on the one side in relation to the benefits to Ministers of having pre-release against the disbenefits to integrity of statistics, my conclusion in that role was that I should withdraw it, which I have done…You asked how it is going; it has gone very smoothly.
Public Administration and Constitutional Affairs Committee, 14 November 2017
He added—
…I am glad to say that the Bank of England has followed suit and the Scottish Parliament are looking at it in relation to Scottish statistics. Of course, there will be some people who are disadvantaged by it who will not have been so happy about it.
Public Administration and Constitutional Affairs Committee, 14 November 2017
Ed Humpherson, Director General of UKSA, told the Committee during our inquiry—
First, why do we care so much about this? It is because, at the heart of what statistics are about, they are a public asset. They are there for public consumption as information that enables the public to understand the nature of the world, the nature of policy and nature of the decisions that are being made. Secondly, that vision of a public asset is underpinned by statistics being equally available to all and not partially available to some audiences but not others, and by their being available as soon as they are ready. Pre-release access seems to run against those two principles.
Economy, Jobs and Fair Work Committee, 7 November 2017
He added—
There is a perception that one set of actors – ministers – gets privileged access that others do not get. Therefore, for the Scottish Government to establish its trustworthiness – to use that term again – it needs to work that much harder to demonstrate the integrity of its production process, and I think that makes its job harder.
Economy, Jobs and Fair Work Committee, 7 November 2017
He was clear no judgement or accusation was being made—
They are genuinely highly professional statisticians who do an excellent job. I just think that pre-release access makes their work harder.
Economy, Jobs and Fair Work Committee, 7 November 2017
As stated in its written submission to this Committee, RSS—
…strongly urges the Committee to call for Scotland’s current rules on pre-release access to be revised. We believe that such privileged access undermines public trust in official statistics as, for example, it creates opportunities for figures to be “spun” to the media or “buried” beneath other announcements.
Royal Statistical Society submission to Inquiry on Economic Data
It said such a reform “would be warmly welcomed by all those committed to statistical integrity”.
Keith Brown, Cabinet Secretary at the time of the Committee’s inquiry, considered “the current arrangements work well” and told us—
…pre-release access to data means that when ministers are called upon to respond quickly to stats at the time of publication, they can do so in an informed way…
Economy, Jobs and Fair Work Committee 14 November 2017
Roger Halliday, the Chief Statistician, explained that the decision on who has advanced sight of statistics in the Scottish Government was his. He felt it was “right” for ONS to end PRA but “not necessarily straightforward” to do so elsewhere.” In his view—
The issue has been overplayed in this debate. There are much more important issues, such as ensuring trust in statistics; data handling to make sure that we treat people’s data securely; setting a culture in which statisticians are independent of influence… and being responsible for recruiting and developing statisticians.
Economy, Jobs and Fair Work Committee 14 November 2017
The current Cabinet Secretary with responsibility for this policy area, Derek Mackay, has maintained his predecessor’s view. He said in a letter to the Committee on 19 October 2018—
Fundamentally this is an issue for the Chief Statistician and…Mr Halliday has made his view on the retention of PRA clear.
Cabinet Secretary for Finance, Economy and Fair Work's letter to Committee, 19 October 2018
In correspondence between UKSA and the Scottish Government more than a decade ago, while the Pre-release Access to Official Statistics (Scotland) Order 2008 was being prepared, UKSA wrote—
Enabling the administration of the day to discuss, and prepare statements about, the statistics whilst not allowing the same access to Parliament or the public is not, in our view, good statistical practice. We believe our view on this is shared by statistical offices and other authorities around the world.
Letter from Minister for Enterprise, Energy and Tourism to the Chair of the UK Statistics Authority, 9 July 2008
It added that international policies in relation to these principles “are mostly clear and supported by both the EU and United Nations” and observed—
The Pre-release Order which has been laid before the Westminster Parliament leaves little doubt that its aim is to reduce the amount of early access currently being granted. The Scottish Order appears to take a different stance and to be focused more on the formalisation and endorsement of pre-release access.
Addressing each point raised in favour of PRA by the Scottish Government at the time, UKSA was of the view the arguments were “less clear-cut than presented”, specifically—
In the normal run of business, statistics should be released without simultaneous accompaniment of ministerial press notices or other statements, as this may “distract attention from the statistical commentary and may encourage public scepticism about the trustworthiness of the statistical product itself”;
“It should normally be possible to correct erroneous data without providing access to the statistics in their final form”, and
The solution to ensuring other publications released at the same time as the statistics (or shortly after) were properly informed would be to publish them when the official statistics become available to all. To issue policy documents at the same time, the UKSA pointed out, “could again be seen by some as distracting attention from, or weakening trust in, the statistics.”
UKSA therefore sought—
…a commitment both to a progressive reduction in the length of time for which privileged access is granted, as well as in the number of officials and Ministers seeing statistics prior to their publication. We would encourage the Scottish Government to adopt statistical policies that promote equal access, the earliest possible publication, and minimise the opportunity to make policy proposals or comments from advance sight of the unpublished statistics.
Letter from UK Statistics Authority to Minister for Enterprise, Energy and Tourism, 29 August 2008
At that time, the Scottish Government described PRA as “necessary and desirable” but acknowledged “we do very much recognise that there are risks associated with pre-release access that need to be controlled”.
The principles of equal access and earliest release possible are integral to the UK Code of Practice, as well as the European Statistics Code of Practice and the United Nations Fundamental Principles of Official Statistics.
These are principles with which both the Scottish Government and the UKSA agree. However, UKSA argues that granting PRA falls short of these principles, is not good statistical practice and risks encouraging public scepticism about the credibility of the statistics themselves.
In putting together the Committee’s data inquiry report there was initially a range of views among members as how best to approach this matter, including: adoption of a default position of no PRA except for exceptional circumstances; removing it for those statistics deemed to be of particular national significance (such as GDP and GERS); or seeking a definitive end to the practice for all Scottish economic statistics.
Ultimately the Committee recommended—
A decade on from the 2008 Order’s introduction, and in light of strong encouragement in favour of ending pre-release from a number of key witnesses to our inquiry, we believe that PRA to economic statistics which are market sensitive – including Scottish GDP, the Retail Sales Index for Scotland (RSIS), Quarterly National Accounts Scotland (QNAS) and Government Expenditure and Revenues (GERS) – should end; and the Scottish Government is invited to set out how it will do so.
That was the prevailing view and recommendation, although a minority of members preferred—
The Committee considers there should be a presumption against pre-release access and invites the Scottish Government to put forward arguments why pre-release access should be continued for specific statistics.
However, even this approach sought – and was premised upon – a presumption against PRA.
There are three strands to what the Committee is proposing—
Removal of PRA entirely for two specific categories of economic statistics;
A phased approach to – and independent review of the impact of – that removal, and
Reducing from five working days to one working day the PRA for those economic statistics where five working days is currently the maximum.
This would remove PRA from two of the four specific economic stats covered by the majority recommendation – the two we wish to address being GDP and Retail Sales – on the basis that neither of these is covered by PRA at the UK level (both being ONS statistics and ONS no longer applying PRA to its own data).
The rationale is that this would represent a compromise between the majority and minority view of the Committee in terms of addressing half of what the majority recommendation called for while not “taking away” anything from the Scottish Government that the UK Government would retain.
Without getting into the detail at this stage of how this strand would be addressed on the face of a Bill, we note that the 2008 Order is made under Section 11 of the Statistics and Registration Service Act 2007, which makes it clear that the rules and principles that may be made in such an Order include “the circumstances in which, or descriptions of statistics in relation to which, pre-release access may or may not be granted”.
This would reduce the one working day PRA covering those two specific economic statistics covered by the majority recommendation – GDP and retail sales – to half a working day after one year of the bill coming into force, removing it entirely after two years, and commissioning an independent review (which could be carried out by an organisation such as the Office for Statistics Regulation) after three years.
The purpose of the review would be to assess the impact of the reduction and then the removal of PRA over that period. The review would be supplemented with a requirement to lay the outcome of the review before the Parliament and a duty on Ministers to respond to the findings of the review either in the form of a publication or by a statement to the Parliament.
The rationale is that this would offer a gradualist approach, built-in review and added accountability.
This would reduce from five working days to one working day the PRA for those economic statistics where five working days is currently the maximum – those being exports, productivity, local government finance, and non-domestic rates. Five days was described by the RSS during our inquiry thus—
…in that regard Scotland is very much an anomaly relative to almost the whole developed world.
Economy, Jobs and Fair Work Committee 26 September 2017
The rationale is that, although this strand does not address those categories of economic statistics specified by the Committee in its majority recommendation on PRA, it would remove a timescale much of the UK statistical community consider untenable; and deliver a change in favour of statistical best practice over what might be termed ministerial advantage – the two competing interests at the heart of the PRA debate.
In the interests of clarity, it might be also helpful to set out what the Bill will not do—
It will not remove anything from the Scottish Government or its ministers that the UK Government or specific Whitehall departments retain;
It will not make any stipulations on any sets of data other than those categories of economic statistics specified. It will not, for example, address health or education statistics, and
It will not call into question the integrity or professionalism of Scottish Government statisticians or other civil servants working in the field of economic or other areas of data.
The remit of the Committee’s data inquiry was: to examine the accuracy, utility and comprehensibility of Scottish economic statistics; to consider what data is required for effective delivery and scrutiny of policy; and to recommend where any improvements might be made.
How we collect, present and interpret data may seem a niche topic, but it shapes decision making in government, in business, in social policy, in the press, and in the public mind.
Our inquiry report made 29 recommendations for improvements to Scotland’s economic data, the majority of which were accepted. The most significant issue on which no progress has been made is that of PRA.
The Committee believes economic statistics should be a public asset, an aid to understanding those political and macro-economic decisions that affect us all, and as such ought to be available when published on an equal and not a privileged basis.
Recommendation
The Committee recommends that the Parliament agrees to this proposal.i
16th Meeting, 2019 (Session 5), Tuesday 14 May 2019
Economic Data (in private): The Committee considered its follow up work in relation to the economic data inquiry and one decision was proposed and decided upon (by division). See full note below.
Record of division in private
Andy Wightman proposed that the Committee recommend that the Parliament approves the proposal for a committee bill. The Committee agreed, by division, to the proposal.
For 4 (Jackie Baillie, Gordon Lindhurst, Dean Lockhart and Andy Wightman); Against 2 (Colin Beattie and Gordon MacDonald); Abstentions 3 (Angela Constance, John Mason and Tom Mason, Committee Substitute).
17th Meeting, 2019 (Session 5) Tuesday 21 May 2019
Economic Data (in private): The Committee agreed to write to the Cabinet Secretary for Finance, Economy and Fair Work.
18th Meeting, 2019 (Session 5), Tuesday 28 May 2019
Economic Data (in private): The Committee considered its follow up work in relation to the economic data inquiry and two decisions were proposed, one decided upon (by division) and another pre-empted. See full note below.
Record of division in private
Andy Wightman proposed the following as text for a new paragraph after paragraph 11 of the proposal report for a committee bill.
"At that meeting the Committee decided to delay publication of its proposal report whilst seeking further views from the Cabinet Secretary for Finance, Economy and Fair Work. The correspondence can be found here."
The proposal was agreed to by division: For 5 (Jackie Baillie, Jamie Halcro Johnston, Gordon Lindhurst, Dean Lockhart and Andy Wightman); Against 4 (Colin Beattie, Angela Constance, Gordon MacDonald and John Mason); Abstentions 0.
John Mason proposed the following.
"That the Committee revises the report to take account of the Scottish Government’s changed position and further considers an amended report at its next meeting."
The Committee agreed that the proposal was pre-empted by the decision on Andy Wightman’s proposal.