EA/2023/0538 - [2024] UKFTT 001054 (GRC)

The Commissioner and the SFO referred to concepts which are sometimes known as the ‘precedent effect’ and the ‘mosaic effect’. These concepts are central to their position that the Relevant Sections were engaged in respect of the Requested Information. These concepts are perhaps best put into context for current purposes by quoting directly from the SFO’s response to the appeal:

“The SFO contends that disclosure of the requested information in this appeal would create a precedent for future FOIA requests, leading to a situation where the SFO would be required to release the costs of other cases, or would at the very least be highly likely to be required to release the costs of other cases (“the precedent effect”).

The SFO contends that the consequence of cumulative disclosures would be to enable individuals, including suspects and defendants in the SFO’s investigations and prosecutions, to construct detailed models of the SFO’s work and the level of resources which the SFO allocates to any given case or category of case, causing direct harm to the SFO’s ability to prosecute economic crime and protect the UK economy, in contradiction of the protections that [the Relevant Sections] are intended to provide (“the mosaic effect”).”.

Mr Shellum’s written skeleton argument referred to part of the Commissioner’s guidance on the ‘mosaic effect’ (which the Commissioner published in respect of section 31 regarding the application of the prejudice test). The guidance we were referred to explains the concept of the ‘mosaic effect’ but then goes on to make two specific points. The first such point is that if a public authority complies with one request for information under FOIA, it can make it more difficult for the public authority to refuse requests for similar information in the future. The second point made is that (in essence) a public authority can take into account, when assessing prejudice, the combined effect of both: (a) the release of information which is currently requested; and (b) the release of information which the public authority could subsequently be required to provide, if the current request was complied with.

The first of those points is essentially referring to the ‘precedent effect’, which we address later below.

The second of those points is, in our view, an over-simplification of the ‘mosaic effect’ (at least in isolation, without reference to the context within which it is to be applied). We say this because it is still necessary for the prejudice test to be applied, such that there must be some causative link between the potential disclosure of the relevant information and the prejudice in question. The prejudice must also be real, actual or of substance and it must relate to the interests protected by the exemption. Further, it must be established that the applicable prejudice “would” or “would be likely to” occur – meaning that the prejudice in question is more probable than not or that there is a real and significant risk of it happening. Therefore the application of, or reliance on, the ’mosaic effect’ is of relevance only insofar as it can demonstrate, or corroborate, that the prejudice in question “would” or “would be likely to” occur.

Further, we also consider that that second point could be misleading by referring to the release of information which the public authority could subsequently be required to provide. This is because it appears to be based on the premise that the public authority might be obliged to disclose further information in the future, without recognition of the potential exemptions to disclosure which could be applicable. Putting this another way, it could be construed as meaning that a public authority cannot take account of applicable exemptions in respect of future requests for disclosure of relevant information.

That second point is, to an extent, a reflection of the ‘precedent effect’ argument we have referred to and which we address later below. We accept that previous disclosures of information may make it more difficult for a public authority to argue that a later disclosure of relevant information is likely to cause prejudice, particularly if there is no evidence of harm (prejudice) being caused in the past. Consequently, we acknowledge the potential difficulties regarding subsequent requests for similar information, such that a situation may arise where initial disclosures of information did not cause actual prejudice but that the risks of prejudice may increase with more disclosures, to the point where a public authority can then withhold relevant requested information.The fact that complying with one request can make it more difficult to refuse requests for similar information in the future was noted by the Commissioner in paragraph 24 of the Decision Notice. However, that paragraph went on to state that that public authorities can consider any harm which could be caused by combining the requested information with the information a public authority could subsequently be required to provide, if the current request was complied with – a point which has the flaws we have already identified (plus see our comments on the ‘precedent effect’ below).

We also accept (as submitted by Mr Shellum) that the application of the prejudice test is a necessarily speculative exercise, in that there is no actual disclosure of the relevant information at the time when the test is being applied and therefore that there must be an assessment of what would happen, or would be likely to happen, were the information to be disclosed. However, as we have noted, there must still be a causative link between the potential disclosure of the relevant information and the prejudice in question - and the prejudice must be real, actual or of substance, as well as relating to the interests protected by the exemption.

Another, fundamental, aspect of the application of the ‘mosaic effect’ when considering the potential disclosure of information under FOIA is that it needs to take account of other information. As the Commissioner explained in the guidance Mr Shellum referred to: “You can take account of any harm likely to arise if someone pieced together the requested information with other information to form a broader picture”. The relevant point is that the “other information” must be other available information. This is an important factor – if the other information is not available then it follows that it cannot be taken into account. The potential caveat to this is the point which the Commissioner and the SFO made about the obligation to release information in the future, which leads us back to consideration of the ‘precedent effect’, to which we now turn.

The witness stated that the disclosure of the requested costs would cause prejudice by creating a “de facto” precedent for future FOIA requests, resulting in the expectation that the SFO would “be highly likely” to be required to routinely release the costs of its cases into the public domain. Similar points were made by the SFO in its response to the appeal (and were recorded by the Commissioner in the Decision Notice).

We consider that those views are illustrative of the concerns to which we referred in paragraphs 47 and 48. In considering any exemptions under FOIA, the test is not whether disclosure would mean that a public authority may be likely to be obliged to respond to further requests, as there is no such exemption, of course (although this may be relevant for establishing if prejudice would or would be likely to occur for the purposes of a prejudice-based exemption). Consequently, a disclosure in one case does not necessarily set a precedent for future disclosures. There is no basis in law for the SFO to be automatically obliged (or even to be ‘highly likely’ to be obliged) to disclose information relating to the SFO’s costs without consideration, on a case by case basis, of the particular facts and circumstances – which would also include consideration of any potential exemptions and the associated Public Interest Test where applicable. In that regard, the ‘precedent effect’ is a misleading concept. Rather, the focus should be on whether an applicable specific exemption is engaged - in this case, this means whether the matters specified in the Relevant Sections would, or would be likely to, be prejudiced if the Requested Information were to be disclosed. For these reasons, we agree with Ms Al-Yassin that the SFO’s reliance on the ‘precedent effect’ is misconceived.

The position is different with regard to the ‘mosaic effect’ which does have a recognised basis in law to some extent, as we noted in paragraphs 49 and 51 regarding the creation of, or increase in, the risk of applicable prejudice under a relevant exemption in FOIA. Therefore we acknowledge the potential relevance of the ‘mosaic effect’. The difficulty for the SFO is that its arguments regarding the ‘mosaic effect’ are largely premised on the ‘precedent effect’.

For example, the witness stated that releasing case costs would directly reveal how much public funding the SFO had chosen to allocate to each specific case. The illustrative context for that, provided by the witness, was based on the SFO being “compelled” to release costs information in response to multiple FOIA requests. For the reasons we have given, we do not accept that the SFO would be compelled under FOIA to disclose information pursuant to similar requests in the future, even if the Requested Information were to be disclosed. The witness conceded during the closed session that “compelled” was a strong word and considered that perhaps it would have been more appropriate to refer to being ‘ordered’ to disclose the information. However, in our view that does not alter the fundamental premise of the SFO’s arguments that future disclosure would somehow be obligatory without consideration or application of exemptions.

A related point, regarding the ability of the SFO to potentially rely on exemptions under FOIA should there be similar information requests in the future, would be the context of those future requests. For example, a request for information relating to a closed case is likely to be different to a request for information regarding a current (or ‘open’) case, for the purposes of both the applicable prejudice which might be relied on as well as any applicable associated Public Interest Test. The witness referred to third parties being able to recreate, by way of the ‘mosaic effect’, a complete picture of how the SFO conducted its operational work or, at least, how its cases were being resourced. This argument was the main point underpinning the SFO’s position regarding the engagement of the Relevant Sections. We understand the SFO’s concerns but we are not persuaded by that argument, for three main reasons (which we set out under sub-headings for ease of reference).

The first reason is based on the concerns we have already raised about the reliance on the ‘precedent effect’ in support of that argument.

The second reason is linked to our point in the preceding paragraph about the likely difference between the SFO’s current/open cases and those which are closed. If the SFO was faced with a request for information relating to (for example) its costs and resources in respect of a current/open case it is investigating, it is not hard to envisage that certain exemptions (such as the Relevant Sections) could be engaged and that the Public Interest Test could favour the maintenance of the applicable exemptions. Obviously this would depend on the circumstances and other factors, but we are making an illustrative point. In contrast, it is more difficult to see how information in respect of a case which has been closed could be used to inform third parties about what resources would be applied for the SFO’s future work or indeed as to what other cases the SFO may work on in the future.

The witness held a different view regarding the position for closed cases, however. They considered that disclosure of costs information pertaining to closed cases could allow third parties to create a picture of the SFO’s work and the level of resources which it allocates to any given case and that this could “point toward” the details of the investigative tools, techniques and powers available to the SFO. The witness also explained that closed cases are often subject to being re-opened at a later date, including where there has been a ‘Victims Right to Review’ of an original decision to close a case. We accept that closed cases could be subsequently re-opened. However, we find there was no other evidence to support the witness’s view that disclosure of costs information could allow third parties to gain the level of information and insight which the witness asserted. Even if we accepted the witness’s evidence on this issue, something more is required than merely ‘pointing towards’ matters. As we have referred to, the law requires that the applicable prejudice must be ‘real, actual or of substance’ and we find that this has not been established.

We also find (and as was argued by Ms Al-Yassin), that the context of the Unaoil case is a very specific one, as it was an investigation where the SFO was found wanting in various ways. In our view, that serves to distinguish the Request from future information requests which may be made about the SFO’s other investigations. The SFO argued that the Unaoil case may be reopened but, notwithstanding our acceptance that closed cases may be re-opened, it seems to us that that is highly unlikely in respect of the Unaoil case, given the specific circumstances of that case and its associated failings (and particularly given that the convictions which were secured by the SFO were later quashed by the Court of Appeal). Consequently, the particular circumstances of the Unaoil case, which is not an open investigation, is such that we find that there is no weight behind the SFO’s arguments that disclosure of the Requested Information would mean that other information would have to be disclosed in the future. Again, there was also no evidence to support those arguments. In this respect, we agree with the submissions of Ms Al-Yassin that the SFO’s arguments in respect of the engagement of the Relevant Sections are based on extrapolated prejudice from its own expectations or fears, rather than from any evidential basis.

The third reason is related to the nature of the Requested Information and the Withheld Information. There are two component parts to this reason.

First, as we noted in paragraphs 42 and 42, the Request did not seek the level of detail which the witness referred to, but rather certain aggregated information (albeit broken down into segments). Accordingly, there is an inherent flaw in the SFO’s position that the Requested Information (which does not require the focussed costs details as asserted by the witness) is likely to be combined with publicly available information in order to establish that the prejudice specified in the Relevant Sections would, or would be likely to, be caused. That position is also partially based on the SFO’s reliance on the ‘precedent effect’, as the witness’s view about publicly available information included information from future FOIA requests.

Secondly, having assessed the Withheld Information (including exploring it with the SFO in our closed session), we find that it was actually very generic in nature. It was set out under very broad headings, without any underlying detail. We simply could not reconcile the Withheld Information with the prejudice in the Relevant Sections as contended by the SFO, even taking into account its ‘mosaic effect’ arguments. The witness stated (in a closed element of their written statement) that revealing the SFO’s costs would indicate certain specific line items of expenditure which would in turn (as referred to in the open witness statement) reveal certain sensitive information which could be damaging to the SFO’s activities. Whilst we understand the premise of the concerns which the witness outlined, this was not reflected in the evidence before us – particularly with regard to the Withheld Information, which contained no such indication (nor was there any other evidence to support that view with regard to any other information).

During the closed session, we asked the witness about the lack of detail contained in the Withheld Information and they explained that, in essence, based on how matters were recorded, it would be difficult to gather information which went into more detail for the purposes of the distinct limbs of the Request. This demonstrates (and we find) that the Withheld Information did not even contain the level of detail which the witness had asserted in respect of their assessment of the Request (as referred to in paragraph 42). We would go so far as to say that we were surprised by the sparsity of the Withheld Information.

We would also comment on a further aspect of the SFO’s position regarding the engagement of the Relevant Sections. The witness stated that (based on their assessment of the Request) individuals would be able to use the Requested Information regarding costs in combination with other publicly available information to “gain greater understanding of every aspect of how the SFO investigates and prosecutes serious fraud, corruption and bribery” (emphasis added). We find that such contention was not corroborated evidentially, including (as we noted in the preceding paragraph) by the nature of the Withheld Information itself. Likewise, whilst we acknowledge and accept the witness’s evidence that the SFO has received twelve FOIA requests for costs information since January 2023 and that there is a high likelihood that similar requests will be made in the future, we find that there was no evidence to support the SFO’s position generally that the Requested Information could be combined with any other information in the public domain, such that the prejudice in the Relevant Sections is established. This is so even when taking into account the evidence and submissions regarding the SFO’s limited size and the focussed nature of its work.

We would summarily conclude by stating that that the prejudice specified in the Relevant Sections has not been established, including in respect of the lower threshold of prejudice (namely “would be likely to”, rather than “would” prejudice). Rather, the SFO’s position was essentially that a precedent would be created, should the Requested Information be disclosed, whereby responses would have to be given to future similar requests for information – and that the relevant prejudice was caused by the cumulative effect of such future disclosures. The Commissioner recorded, in paragraph 17 of the Decision Notice, that the SFO explained this position as follows:

“The SFO considers that disclosure of the requested information in this case would create a precedent for future FOIA 2000 requests and lead to a situation where the SFO would be required to release the costs of all cases. Section 31 is therefore engaged in this response because of the prejudice or likely prejudice caused by the cumulative effect of disclosing information in response to a series of similar requests (the ‘precedent effect’).”

For the reasons we have given, we find that the SFO’s reliance on the ‘precedent effect’ is flawed and that its position in respect of the prejudice in the Relevant Sections was underpinned by its mischaracterisation of the Request. There was also a lack of evidence to support the SFO’s arguments, which we find were largely based on mere supposition. We accordingly find that SFO and the Commissioner have not established that the disclosure of the Requested Information would, or would be likely to, prejudice the matters specified in the Relevant Sections.

For all of the above reasons, we find that the Relevant Sections were not engaged in respect of the Request.

As we have concluded that none of the Relevant Sections were engaged, it is not necessary for us to go to consider the Public Interest Test.