George Greenwood v The Information Commissioner & Anor

The DHSC relies on the evidence of Mr Cramp and contends that significant weight can be placed on his evidence, particularly in the light of additional work undertaken to review the data. It is worth considering the terms of the request which sought all correspondence between Mr Hancock and Ms Coladangelo using Mr Hancock’s or Ms Coladangelo’s departmental email and any private email that Mr Hancock used for government business.

Mr Cramp confirmed that four departmental email accounts were identified for Mr Hancock. There is one closed DHSC email account for Mr Hancock, two closed DHSC email accounts for Mr Hancock’s private office (private office one and private office two), and one further closed DHSC email account for Mr Hancock. Mr Cramp explains that the fourth account was used solely for teleconferencing and not as a correspondence address and so was considered outside the scope of the first request. This was not disputed, and we accept that it is out of scope.

In his witness statement and in his oral evidence Mr Cramp explained the process used to search for and identify emails falling within the scope of the first request for information. He confirmed that searches were made of the three DHSC email accounts relating to Mr Hancock for emails where Ms Coladangelo’s DHSC email or her Oliver Bonas email was a participant in any field. Mr Cramp explained in his witness statement that any emails found within the private office one account were considered to be out of scope of the request because they were operated by Mr Hancock’s private secretaries and not by Mr Hancock himself. The DHSC therefore considered that they did not fall within the terms of the request which was for all correspondence between Mr Hancock and Ms Coladangelo. We find that this was an error.

The two private office accounts are just that, email accounts for the private office of Mr Hancock. The role of a private secretary is as the main link between a government minister and officials in the department or ministry. For this reason, even if a private office email account is operated by a private secretary, we consider that correspondence with a private office email account ought to be regarded as correspondence with the relevant minister, in this case Mr Hancock. For these reasons, we find that the two private office email accounts are within scope of the first request.

In cross-examination Mr Cramp confirmed that no searches as outlined in paragraph 35 above were made of Ms Coladangelo’s DHSC email account. In other words, no searches were made of Ms Coladangelo’s DHSC account for emails where any of Mr Hancock’s DHSC emails or any of his private emails were participants in any field. We find that this is a material omission, which is reasonably likely to have resulted in a failure to identify all emails within scope of the first request.

There is agreement by all three parties that the decision notice issued on 28 March 2023 is not in accordance with the law and we have been asked to formally reflect that fact in this decision. Section 58 of FOIA governs the determination of appeals. If the tribunal finds that the decision notice was not in accordance with the law, it may allow the appeal or substitute such other notice as could have been served by the Commissioner. Both the appellant and the Commissioner have asked us to substitute the decision.

We find that the decision notice issued on 28 March 2023 was not in accordance with the law. The basis for that finding is that the respondent’s decision that it did not hold the information requested was wrong in law because it failed to consider whether the requested information was held on behalf of the DHSC by Mr Hancock or Ms Coladangelo or any other person included in their communications such that section 3(2)(b) of FOIA might apply. Prior to the decision notice being issued on 23 March 2023, the DHSC confirmed that WhatsApp messages from Mr Hancock’s personal device were held on a computer system on its behalf. The fact that the DHSC did not transfer those messages into its official record following Mr Hancock’s resignation is not relevant for the purposes of section 3(2)(b) of FOIA.

The DHSC accepted that the meaning of correspondence must be considered by conducting a real world analysis, but sought to persuade us that the nature of communication in WhatsApp groups meant that it was not ‘true’ correspondence. With respect, we find that the DHSC’s analysis falls some way short of a real world analysis. In fact, it was an analysis that concluded that unless correspondence consists of one person corresponding directly with another, it is not ‘true’ correspondence. A real world analysis would have recognised that correspondence in the age of multiple methods of electronic communication can take different forms. We find the fact that simply because one or other of the relevant parties did not respond or may not have responded to a particular message does not mean that communications within a WhatsApp group cannot be considered to be correspondence.

We do not consider that any intention to correspond is required for messages in groups to be considered correspondence. Unsolicited correspondence may still be considered to be correspondence and therefore an intention to receive is not required. Equally, a response is not always required to correspondence and it follows that an intention to respond is not required for there to nevertheless be correspondence.

For the same reason, we also reject the submission on behalf of the DHSC that by sending a message to a WhatsApp group a person should be considered to be broadcasting rather than corresponding. We accept that some messages may well be for the purpose of notifying members of a group of a particular matter without a specific requirement for a response. However, the same can be true of emails or other forms of communication such as letters. Furthermore, the fact that an individual does not respond to a particular message is not an indication that the incoming message was not correspondence with them. They may have nothing to say or they may respond in another way, for example in a meeting, by telephone, by email etc. This is conceivable where for example a question may be posed, but the medium of a text message is inappropriate for the answer, i.e. because it would be too long and therefore impractical.

We find that the members of the various groups did not simply ‘happen’ to be members. We find that they were members of different groups for specific purposes. For example, the group Matt and his NEDs was a group consisting of Mr Hancock and two of his non-executive directors. We find that this was a channel of communication set up for the express purpose of the three individuals communicating with each other.

We reject the distinction identified at paragraph 27 of the skeleton argument on behalf of the DHSC between email and WhatsApp messages. When drafting an email a sender may make a choice as to which address(es) they include in the ‘to’, ‘cc’ and ‘bcc’ fields, particularly if they are initiating the email correspondence. A WhatsApp group has a specific and deliberate membership which can be seen on the Group Info screen where members of the group are listed. The default is that a message is seen by all members of the group when sending a message in a group chat. Similarly, WhatsApp also provides the option of replying privately or messaging a particular member rather than the entire group. In our view therefore, by sending a message or replying to a message in a WhatsApp group without choosing to reply privately or to message a particular member, a person is making a deliberate choice to communicate with all the members of the group.

The appellant’s second request sought all correspondence ‘relating to government business’. In his witness statement Mr Harris states that the DHSC categorised the WhatsApp messages as follows:

The DHSC concluded that the messages falling into categories (b) to (d) were not within scope of the second request because they did not relate to government business. Specifically, the DHSC asserted that the correspondence did not concern policy advice given by Ms Coladangelo to Mr Hancock, nor any other input she may have had in departmental business. We disagree. We find that the WhatsApp groups were used for the purposes of carrying out government business. While there may have been some personal messages which would not be disclosable, we find that the purpose of various groups, and in particular Matt & his NEDs, was as a forum for discussing government business. We explain our reasons for this finding in the CLOSED annex.

For all of these reasons, we find that by applying a broad, real world interpretation of ‘correspondence between’, messages sent and received by Mr Hancock and Ms Coladangelo in WhatsApp groups of which they are members is correspondence between them and accordingly they are within scope of the second request.

The DHSC asserts that the second request is vexatious because of the resources that will be required in order to respond to that request. While it is not material for the purposes of considering this appeal, we note that the DHSC made a deliberate decision to rely on section 14 of FOIA rather than section 12, which provides for an exemption where the cost of compliance exceeds the appropriate limit.

The term ‘vexatious’ is not defined in FOIA. We must therefore consider relevant caselaw to see how the term has been interpreted for the purposes of section 14.

In Dransfield v IC and Devon CC [2015] EWCA Civ 454, [2015] 1 WLR 5316 the Court of Appeal held that a comprehensive or exhaustive definition of vexatious is not appropriate. It held that the emphasis should be on an objective standard and, consistent with the constitutional nature of the right in question, public authorities face a high hurdle in establishing that a request is vexatious. The decision maker is required to consider all the relevant circumstances in order to reach a balanced conclusion as to whether a request is vexatious.

Other principles that can be identified from the caselaw include:

We do not understand it to be in dispute that there is significant public interest in the scrutiny of how taxpayers’ money is spent; in ensuring that government is accountable for achieving value for money and that the process is transparent, as well as exposing any potential conflicts of interest which may have an impact on matters of government and in understanding how the business of government was being carried out. In the context of events around the time of the appellant’s request, we consider that there is significant public interest in disclosure of the requested information.

The DHCS identifies two bases on which the appellant’s second request is vexatious. The first is the burden it would impose and the second is that the exercise of reviewing and disclosing messages would have limited value or purpose. This is because the posts have already been disclosed in the course of the COVID-19 Inquiry and were leaked by a journalist to the national press. The DHSC submits therefore that any public interest in the messages is likely to have been satisfied in the course of the Inquiry and/or the public reporting.

We find that in the light of the number of WhatsApp messages potentially in scope of the second request, it is arguable that the request would impose a substantial burden on the DHSC; however, we do not accept that the DHSC has in fact shown that this would be the case.

We note that we heard submissions in the CLOSED session in respect of this process, but they were repeated in the OPEN session and therefore we consider that it is appropriate to include them in the OPEN reasons.

The DHSC relies on the evidence of Mr Harris to demonstrate why the second request creates a substantial burden such that it should be considered vexatious. Mr Harris explains in his statement that the DHSC identified eight WhatsApp group chats which are potentially within scope of the appellant’s request. The DHSC identified one of those groups – ‘MSM’ which was used to provide the estimate of how long it would take to comply with the request. This was a group chat with a total of 11,008 posts in total, i.e. from all 35 members of the group. This was the group with by far the largest number of posts, the next largest had a total of 3,856 posts and the smallest had a total of 62 posts.

Mr Hancock was said to be responsible for 28% of the messages in MSM and Ms Coladangelo was responsible for 1% of the messages, giving a combined total of 28%, or 3,192 messages.

In his statement, Mr Harris identifies seven steps he says the DHSC would need to take in complying with the request:

In his oral evidence Mr Harris confirmed that he did not complete any of the searches himself, but discussed with the team which did. He also confirmed that he had not provided an explanation of how the DHSC’s estimate of the time it would take to carry out the above steps in relation to MSM was calculated but that he had discussed it with the team. He stated that the estimate of 528 hours was calculated on the basis of all 11,008 messages and not the 3,192 originating from Mr Hancock and Ms Coladangelo.

Although Mr Harris accepted that the estimate of 528 hours work was based on all 11,008 messages, he stated that the estimate would still be over 300 hours. He accepted that there was no detailed analysis of how that estimate was reached and gave an explanation in his oral evidence. He accepted that he had not seen the messages but understood that some were lengthy and contained attachments whereas others were simply a single emoji. He said that he had discussed the matter with the team which gave an estimate of 2.5 minutes per message. He indicated that the team had considerable experience and had dealt with many requests, and so he considered the estimate was reliable.

Mr Harris gave evidence explaining why he considered all of the steps he identified in his witness statement were necessary. He said that deciding whether particular messages are in scope of the request would be difficult, particularly determining whether they related to work or not. Mr Harris rejected the idea that obtaining comment from Mr Hancock and Ms Coladaneglo would be as simple as ‘hitting send’ and asking them for their views. He pointed to the likelihood of subsequent exchanges between them and did not consider it would be as light touch a process as was suggested.

Mr Harris accepted that any discussions with DHSC about whether messages were in scope would be in relation to those where it is not obvious, in other words, a subset of the 3,192 messages. Mr Harris confirmed that redacting messages would be straightforward and that preparing relevant spreadsheets was also straightforward.

When asked whether steps (b) and (f) were in fact duplication, Mr Harris considered that (b) was broader and that (f) was undertaken once a decision had been made about what was disclosable and sharing that with the relevant parties to get their views. When asked why it would not be possible to filter the messages and make the decision as to what is disclosable before sending them to the parties for comment, Mr Harris was unsure, but surmised that it may be due to the code of practice which stipulates the two stage approach. He also considered that even if steps (b) and (f) did amount to duplication, and only one were required, the time it would take to comply with the request would still be in excess of what is reasonable.

In relation to the final step of checking for errors, Mr Harris stated that this may be complex and require a second pair of eyes and which would take time.

Having considered the evidence, we do not accept the second respondent’s assessment of the time it will take to comply with the second request. Although the respondent chose the WhatsApp group with the largest number of messages to review to provide the estimate, there is nothing to suggest that this is representative of the other groups. All but one of the other groups contains far fewer members and far fewer messages.

Although we accept that the messages vary in length and content, we consider that the longer, more detailed messages and ones with attachments are likely to be more easily identified as being in or out of scope because the content and its relevance will be more obvious. Similarly, very short messages including ones which contain only emojis will be easily included or excluded. The process of deciding whether the majority of messages are in or out of scope is likely to be straightforward. We accept that there will be some messages where scope is not obvious, but we consider that this will be a minority of the relevant messages.

We agree with the appellant that steps (b) and (f) are duplicative. There appears to be no reason why two separate consultations with the authors of the messages are required. Mr Harris was unable to confirm whether the time taken by the relevant parties to consider the messages was included in the 2.5 minute per message estimate, but believed that it was. We find that this is an error, because although the second respondent may be obliged to seek their comment, the time they take to review their own messages is a matter for them and cannot contribute to whether the burden on the DHSC is unreasonable.

For these reasons, we find that the DHSC has failed to demonstrate that the second request is in its entirety vexatious solely on the basis of the burden it creates.

The appellant argues that the DHSC had a duty to assist him to narrow his request. He relies on section 16 of FOIA and the Commissioner’s guidance in respect of that section. Section 16(1) imposes a duty on a public authority to provide advice and assistance to a person who has made a request for information. This is not an open-ended duty, requiring the public authority to assist ‘so far as it would be reasonable to expect the authority to do so’. In that sense it does not create a requirement that the public authority must have offered advice and assistance to narrow a request before concluding that the request is vexatious on the grounds of the burden it would create. We accept of course that offering advice and assistance is good practice as confirmed in the Commissioner’s guidance. We accept the Commissioner’s submission that there is no statutory duty on a public authority to offer advice and assistance under section 16 before they may rely on section 14.

This is not a case where the appellant’s request not sufficiently clear to describe the information sought. Similarly, this is not a case where clarification was needed to try and bring the request within the cost limit in section 12 of FOIA. The DHSC did not rely on section 12. In this case the DHSC’s position is essentially that the request would be vexatious due to the burden it creates regardless of whether the appellant was able to narrow it. Therefore, even had the DHSC offered advice and assistance to the appellant, we find that the outcome would not have been materially different.

Our view is reinforced by subsequent events. On 14 January 2024 the appellant indicated that he was willing to narrow his request to two WhatsApp groups, namely ‘Matt & his NEDs’ and ‘Level 4’. The DHSC responded, indicating that even in light of the narrowed scope, it considered the burden was still such that the request remained vexatious. We do not accept the DHSC’s conclusion and find that the revised request is not vexatious solely on the grounds of the burden it creates.

We have had regard to the fact that the WhatsApp messages have already been disclosed in the course of the COVID-19 Inquiry. The appellant did not dispute that this was the case. We bear in mind that the Court of Appeal held that the possible availability of the information through disclosure in other proceedings is irrelevant to the question of the seriousness of the purpose for which the information is sought. We are satisfied that the appellant’s purpose in seeking the information was a serious one and that there is a reasonable foundation for thinking that the information sought would be of value to the appellant in his role as a journalist and to the public.

We have considered the fact that the information is in the public domain. The reporting of the leaked messages began in or about 1 or 2 March 2023. This is after the DHSC’s responses to the second request and before the Commissioner’s decision in respect of the second request. We find that the fact the messages were leaked is not relevant to the question of whether the appellant’s request is vexatious. It is however potentially relevant to what steps we might require the DHSC to take.

Having considered the evidence and submissions, for the reasons set out above, we find that none of the factors either singly or in combination establish on the balance of probabilities that the appellant’s second request is vexatious.

In conclusion we find that the DHSC holds the requested information and that the DHSC must comply with the second request, save that the request for WhatsApp messages is limited to messages sent and received in the two groups in the appellant’s revised request, i.e. Matt & his NEDs and Level 4.

The appeals are allowed in the terms set out above.