Although the report's publication led to renewed calls for a Freedom of Information (FOI) Act, this reform is not mentioned in the report itself - apart from a passing reference in the evidence of the former foreign secretary, Lord Howe, in which he suggests that even the American FOI Act would not have produced any greater openness about arms sales than occurred in the UK. (1) 

However, we believe that an effective Freedom of Information Act would make a significant contribution to the problems identified in the report. Above all it would signify a substantial change of culture. The present culture was perhaps most tellingly illustrated by Lord Howe, when he told the Scott inquiry that:

"...there is nothing necessarily open to criticism in incompatibility between policy and public presentation of policy...It [ie the Government] is not necessarily to be criticised for difference between policy and public presentation of policy" (2) 

We believe that statement would come as a considerable shock to most people. It goes considerably beyond suggesting that it is sometimes not possible to reveal all the facts about a matter, and asserts that there is nothing objectionable in the government giving an account of its actions which is so divorced from the actual position as to be "incompatible" with it.

A Freedom of Information Act would go some way towards asserting the importance of the public's right to know. It will not produce absolute openness, since a considerable list of exemptions are likely to be needed, but it will shift the norm in favour of greater disclosure.

Such legislation would broadly influence the culture of government in several ways. First, it would provide a legally enforceable right to official information which could be used to challenge and override any unjustified withholding. Second, it will encourage the public expectation of openness, so that instead of a resigned acceptance of the inevitability of being kept in the dark, people will be more likely to ask for information and to challenge refusals. Third, it will permit a more positive response on the part of many officials who may be ready, or even prefer, to operate in a more open manner but are inhibited from doing so by existing conventions. Fourth, it may persuade departments to volunteer information, in the knowledge that they would be forced to release it anyway if the information is formally requested under the legislation.

The Code of Practice

To what extent may such changes already be occurring as a result of the introduction of the Code of Practice on Access to Government Information? In the British context, the Code is a significant move in the right direction. It provides, for the first time, an opportunity to have an independent review by the Parliamentary Ombudsman of the decision of an official, or even a minister, to withhold information. Although the Code contains numerous broad exemptions from the obligation to disclose, it permits many kinds of exempt information to be disclosed where there is an overriding public interest in openness - a positive provision. In the Campaign's experience the Code can sometimes be effective in eliciting previously withheld information from government in certain areas. On the other hand, the Code operates slowly, its procedures are cumbersome, and it demands what for many applicants will be an unrealistic degree of determination and patience.

The government asserts that the Code represents a "is less legalistic, less rigid and less costly" approach to disclosure than a Freedom of Information Act. (3)   Yet it is also less effective. The length of time taken for the Ombudsman to complete his open government inquiries doubled to an average of 32 weeks in 1995 (4)   - considerably longer than the average time taken under Canada's FOI legislation. (5)   At the time of writing, one complaint by the Campaign for Freedom of Information to the Ombudsman has so far been under investigation for 12 months while another has lasted 14 months, without either yet reaching a conclusion.

The Ombudsman's 1995 annual report indicates considerable resistance to change in some quarters:

"there is a tendency in some departments to use every argument that can be mounted, whether legally-based, Code-based or at times simply obstructive, to help justify a past decision that a particular document or piece of information should not be released instead of reappraising the matter in the light of the Code with an open mind. I have found it time-consuming to have to consider a whole series of different defences, even when many of them prove to have no real foundation. That is one reason why it has taken longer than I would wish to complete my investigations during 1995." (6) 

Another factor which has undermined any potential impact of the Code on traditional Whitehall culture has been the lack of government publicity for it - a point on which the recent report of the Select Committee on the Parliamentary Commissioner for Administration expressed criticism. (7)   Although some welcome national press advertising for the Code has been undertaken in recent weeks, it received minimal publicity during the first two years of its operation. (8)   As a result, relatively few requests have been made under it, and there have been so few Ombudsman investigations - only sixteen were begun in 1995 (9)   - that no great impact on Whitehall norms could be expected. Indeed, we understand that because of the low level of complaints, some of the new staff taken on by the Ombudsman's office to handle the anticipated level of open government work have since been reallocated to conventional maladministration cases.

However, the Code's shortcomings are not limited to lack of publicity. Because it is not legally enforceable it indicates that government can, as a last resort, refuse to comply with a recommendation of the Parliamentary Ombudsman in relation to disclosure. Even if this proves to be a rare event, it signifies that the government will ultimately have the last word - not the basis for a fundamental change in culture.

In most circumstances the government will be reluctant to ignore the recommendations of the Ombudsman. However, where a disclosure may lead to substantial political embarrassment or undermine a key policy, ministers may decide that to be criticised for ignoring the Ombudsman's recommendations is the lesser evil. For some time the government was prepared to reject the Ombudsman's recommendations on the payment of compensation to the victims of blight caused by the Channel tunnel rail link. That it subsequently modified its position may owe as much the nearness of a general election and to a ministerial reshuffle (and a new Transport Secretary who could reverse the decision without loss of face) as to the merits of the case.

More detailed comments on the operation of the Code are contained in the Campaign's evidence to the Select Committee on the PCA, and in the Committee's own substantial series of recommendations. In particular, we strongly welcome the Committee's recommendation that there should be a Freedom of Information Act.

The Matrix Churchill case

Would a Freedom of Information Act have made any difference in relation to the arms to Iraq affair? Ministers have argued that it would not, suggesting that most of the relevant information would have been protected under the various exemptions in any such legislation, for example relating to policy advice, commercial confidentiality and the work of the security services. (10) 

We believe that FOI legislation would have made a difference, though whether it would have been so substantial as to have prevented the Matrix Churchill affair altogether is another matter. The Scott report identifies many examples of information which was withheld without good reason, and which would presumably have had to be disclosed under any reasonably effective FOI Act. These include a list of the arms related exports to Iran, disclosure of which was refused even though the official concerned acknowledged it could have been disclosed in aggregate form and despite the fact that Sir Charles Powell, the prime minister's foreign policy adviser, was in favour of its disclosure; (11)   the number of export licenses granted to both Iraq and Iran for chemical warfare defensive equipment (12)   and the size of the credits granted by the Export Credits Guarantee Department to the individual countries. (13)   These and other items of information could have been obtained under an FOI Act, perhaps even by Members of Parliament themselves.

Even more sensitive, exempt, information could in theory also be disclosable under the "public interest override" provisions that might be expected in such an Act. The Right to Know Bill, introduced by Mark Fisher MP in 1993, allowed exempt information to be disclosed if there had been wrongdoing and if in the circumstances the potential harm from the disclosure was outweighed bythe public interest in openness. (14)   A comparable public interest override appears in the Code of Practice on Access to Government Information. (15)   Such provisions would mean that, in circumstances such as those which led to the setting up of the Scott inquiry - where a clear public interest justifying exceptional disclosure had been acknowledged - much normally confidential information would become available.

Finally, a Freedom of Information Act might provide for a public interest defence to charges under the Official Secrets Act 1989 and protection against disciplinary action for civil servants who disclose information in the public interest in particular circumstances.

Policy advice

A central question under any FOI Act is the degree to which civil service policy advice might be disclosed. At present, the broad class of 'advice' is withheld from the public and from select committees. If it is sought during legal proceedings, it becomes subject to a claim for public interest immunity (PII). Where such a claim is made, the final decision is taken by the judge after considering whether the case for confidentiality is outweighed by the case for disclosure in the interests of justice. To the extent that the court weighs the two competing interests, a balancing exercise takes place which does not occur when disclosure of advice is sought by the public or Parliament.

But the case against revealing advice is in essence the same, whether it is raised in relation to disclosure to Parliament or on the face of a PII certificate. These are essentially that disclosure of advice, however innocuous in itself, would undermine the frankness and candour of future internal deliberations; and that publicity would undermine the decision making process, for example, by prematurely exposing preliminary ideas to criticism.

The Scott Report is sceptical as whether these arguments justify protecting the entire class of advice, regardless of the sensitivity of actual documents concerned. Sir Richard concluded: "I find it difficult to accept that these are satisfactory grounds for a class claim in the first place." (16)   Although Sir Richard was referring to the class claim advanced in relation to PII during criminal proceedings, it may be reasonable to suggest, that this particular comment applied to the justification advanced for protecting the entire class from disclosure in other circumstances also.

Sir Richard's analysis of the legal rulings on PII illustrate a strand of judicial scepticism about some of the arguments against the disclosure of policy advice. That may be reinforced by the Chancellor of the Exchequer's decision in April 1994 to publish the minutes of his monthly meetings with the Governor of the Bank of England, precisely the kind of advice whose disclosure had previously been most fiercely resisted.

A British Freedom of Information Act will inevitably include an exemption for policy advice. The question is how broadly such an exemption should be drafted. The existing conventions sometimes appear to have been drawn up in order to protect the most sensitive Cabinet deliberations, yet then been applied with equal vigour to the innumerable lesser classes of internal communications, some of which (such as those in which a technical opinion is offered on the interpretation of data) are difficult to regard as policy advice at all.

The Campaign has previously proposed that any exemption for policy advice should be drafted in such a way as to distinguish between the recommendations of an individual, on the one hand and, on the other, the analysis and interpretation of factual data, and projections based on such data. We have also proposed that "expert advice" in the sense of a specialist professional or technical opinion, should also be excluded from the scope of any police advice exemption. (17)   This envisages that the public should be entitled to have access to the government's best estimates of the likely effects of a new policy - and that to withhold such information on the grounds that it can be regarded as advice, undermines accountability.

The Scott Report suggests an alternative approach, which is essentially to abandon policy advice as a "class" exemption and allow such material to be withheld on a "contents" basis only. This implies that advice could be withheld only where it was shown that to disclose it might be harmful to the decision making process.

It now appears that we may already have reached this point under the Code of Practice on Access to Government Information. Exemption 2 in the Code applies to "Internal discussion and advice" and exempts:

"Information whose disclosure would harm the frankness and candour of internal discussion, including:

• proceedings of Cabinet and Cabinet committees;

• internal opinion, advice, recommendation, consultation and deliberation;

• projections and assumptions relating to internal policy analysis; analysis of alternative policy options and information relating to rejected policy options;

• confidential communications between departments, public bodies and regulatory bodies."

At first sight this looks like a comprehensive "class exemption" covering all material falling within the four broad sub-headings. This is certainly what seems to have been intended, judging from the Office of Public Service's guidance on the Code. This states that one reason for protecting advice is to prevent differences of opinion between officials or between officials and ministers being revealed as this could:

"prejudice working relationships and effective discussion of policy, especially if it led to dissenting or different views being quoted in political argument to attack the policy'

Yet it immediately adds:

"It is not the intention, however, to withhold this class of information only where internal differences and disagreements would be revealed" (18) 

These and other passages suggest that the exemption is envisaged as applying to the entire class of advice (indeed the above passage refers to the "class of information") regardless of whether harm would result from a particular disclosure. However, the exemption itself refers to disclosures which would "harm the frankness and candour of internal discussion". This is surely a harm test (indeed it has been described as such by the Chancellor of the Duchy of Lancaster, Roger Freeman (19) ). The implication is that requests for policy advice under the Code may succeed where it can be shown that a particular disclosure can be made without harming frankness.