Jonathan Bourne v The Information Commissioner & Anor

In our view this is a clearly drafted and unambiguous request. Mr. Bourne explicitly asks for the title/UPRN look up table sold as part of the National Polygon dataset:

“I would like the title UPRN lookup table currently sold as part of the National Polygon Dataset to be released under the freedom of information act.”

He asks in the alternative for a section of that information, namely the look up table for title numbers and UPRN for Land Registry data already in the public domain:

It is not a request for the Land Registry to provide the UPRN of the properties in their Price Paid dataset. Mr. Bourne states that he has not changed his position. He states that throughout his request, it has been clear that what he is requesting is the ability to match the Price Paid data with other government datasets.

Under FOIA, a requestor can only request information, and the tribunal can only order a public authority to provide information. We cannot order a public authority to ‘add the UPRN to the OCOD dataset and the Price Paid Dataset’. We cannot order a public authority to create a document or a dataset that does not already exist.

We accept that Mr. Bourne’s intention was to be able to match Land Registry datasets, such as the Price Paid dataset, with other government datasets, and that for this he needed the UPRN of the properties on the Land Registry datasets. The information he chose to request in order to achieve that was the look up table for title numbers and UPRN. It is the disclosure of that information in relation to which we must assess the likely prejudice. We are not assessing the likely prejudice of the Land Registry adding UPRN numbers to its existing databases.

The difficulty is not that Mr. Bourne has ‘changed his argument’. The difficulty is that our jurisdiction is limited to considering whether the information that was requested from the Land Registry should have been disclosed or not. We are not entitled to consider the disclosure of any other information specified in the grounds of appeal or otherwise. This is not simply a ‘technicality’. It is the whole basis of the FOIA regime. That is why Mr Bourne is not permitted to depart from the scope of his original request at this stage.

We note Mr. Bourne’s complaint that, in contrast, the Land Registry was allowed to change their argument during the course of the Commissioner’s investigation. The short answer is that there is binding authority that confirms that the public authority is allowed, as of right, to raise new exemptions even before the tribunal (see for example Birkett v Defra [2011] EWCA Civ 1606). The longer answer is set out in that decision, should Mr. Bourne be interested in the rationale.

As the first-tier tribunal in Hogan v Oxford City Council and the Information Commissioner (EA/2005/0026, EA/2005/0030) observed at para 27, where the specified activity or interest which would be likely to be prejudiced is a public interest, like the prevention or detection of crime, there is an obvious overlap between whether or not the section is engaged and any subsequent application of the public interest test. We have therefore combined our consideration of the two issues, but we bear in mind that although the relevant factors may overlap, the questions that we have to answer are different.

The applicable interest in this case is the prevention and detection of crime. It is important to note that s 31(1)(a) is engaged where there would be likely to be prejudice to the prevention and detection of crime. It does not require the public authority to show that disclosure will lead to an increase in crime.

When deciding if the section is engaged, we must decide if the public authority has satisfied the evidential burden of showing that some causal relationship exists between the prejudice being claimed and the potential disclosure; if the prejudice is real, actual or substantial; and whether the chance of prejudice is more than a hypothetical or remote possibility i.e. is there a real and significant risk of prejudice?

We accept that the the use of a licence agreement to access the requested information allows the Land Registry to have full visibility of who is accessing information about the titles registered with them. We accept that this is an essential aspect of how the Land Registry identifies or minimises fraudulent activity. We accept that the licencing of this dataset is a key tool for preventing, detecting and investigating registered title fraud, because we accept that the information could be used by criminals or fraudsters to target properties to commit registered title fraud.

In our view it is not necessary for the Land Registry to identify the precise mechanism by which criminals or fraudsters could use the information to target properties to commit registered title fraud. We acknowledge that UPRN information is freely available, and title numbers are available to purchase on an individual basis, but this information is not otherwise available as a single combined dataset outside of the National Polygon Service.

As a matter of common sense, it is likely that uncontrolled access to a single dataset of title numbers of 26 million registered properties, along with the associated UPRN allowing data to be matched with other publicly available data, would make registered title fraud easier to commit.

We are persuaded on this basis that there is a causal relationship between the unlicensed release of the look up table for UPRN/title numbers and prejudice to the Land Registry’s ability to prevent and detect crime. We find that this prejudice is real, actual and substantial and that the chance of prejudice is more than a hypothetical or remote possibility.

Mr. Bourne’s arguments questioning whether the Land Registry is, in reality, concerned about fraud, focus on the effect of adding UPRN reference numbers to its databases. The Land Registry makes clear that they do not consider that that, in itself, creates a risk of title fraud. There is already a publicly available dataset which adds the UPRN to the Price Paid dataset. The risk arises from the unlicensed disclosure of a dataset containing title numbers.

For the reasons set out above, we find that the release of the list would be likely to prejudice the prevention and detection of crime. We find that section 31(1)(a) is engaged.

Turning to the public interest, taking into account our conclusions set out above, we find that there is a very significant public interest in maintaining the exemption. We accept that the release of this information would make it easier for a fraudster to commit registered title fraud, which would lead to losses from the public purse as a result of the indemnity. It would be more difficult for the Land Registry to prevent and detect this crime, because access to the dataset would not be controlled by a licence giving them extensive monitoring powers. We find that this risks undermining the integrity of the land register which is part of the United Kingdom’s Critical National Infrastructure.

Turning to the public interest in disclosure, the government guidance, which is part of the open standards mandated for government use, states that UPRNs must be used. The Land Registry is committed to improving its datasets in the coming years and to releasing more data, supported by UPRNs where resource and funding allows. Releasing this information on an open government license would enable UPRNs to be added by the public now to at least some of the datasets published by the Land Registry.

As the Land Registry have pointed out, as the Price Paid dataset does not include title numbers, Mr. Bourne would not be able to use the requested look up table to find out the UPRN numbers of the properties in the Price Paid dataset. Further, there is already a published dataset (and coding) linking UPRNs to Price Paid data through a study by Urban Big Data Centre (UBDC) at the University of Glasgow (which was funded by Ordnance Survey). This reduces some of the claimed public interest in disclosure.

The tribunal understands that what used to be referred to as the ‘OCOD’ dataset, mentioned in the outcome section of the grounds of appeal, does include title numbers, and therefore we assume that the requested look up table would enable Mr. Bourne to find out the UPRN numbers of the properties in the ‘OCOD’ dataset.

We accept that there are likely to be benefits for researchers and members of the public in terms of enabling research and comparisons of data without having to pay the £20,000 annual fee for accessing the data. We accept that research in this area is likely to be important for the reasons set out in Mr. Bourne’s final submissions. Real harm is caused by housing difficulties, and policies and pressure for change need to be supported by evidence.

Further, there is a general interest in transparency in relation to the land register, given its importance and status as part of the Critical National Infrastructure.

We note that research can be undertaken by those with access to significant resources, because the dataset is available under licence. Given the level of the annual charge, we do not place very significant weight on this in reducing the public interest in disclosure.

For all those reasons we conclude that there is a reasonable level of public interest in disclosure, but we conclude that this interest is outweighed by the very significant public interest in maintaining the exemption.

For the reasons set out above the appeal is dismissed.

Both parties had expressed a willingness to engage in informal discussions. Now that the tribunal process has concluded, we hope that the parties will feel able to commence that process.