Post: 28 April 2017
Land Registry’s Price Paid Data (PPD) is a large dataset containing address-level records of residential property transactions in England and Wales from 1995 to the present.
In 2004 Land Registry began to offer Price Paid Data for re-use on commercial terms. In 2012, as part of a wider open data initiative led by the Cabinet Office, Land Registry began a phased release of the PPD under the terms of the Open Government Licence. The full dataset is now freely available and updated monthly.
Land Registry has always maintained that the Price Paid Data is not personal data. It’s 2013 Privacy Impact Assessment review, prepared in support of the open data release, included the following:
We also spoke to the Information Commissioner’s Office (‘ICO’) and confirmed the steps and ongoing evaluation we have undertaken over the last twelve months. We confirmed our view that monthly PPI was not personal data but property related information remained unchanged.
I’ve previously expressed misgivings about the non-personal status of Price Paid Data. Land Registry’s position seemed to be inconsistent with the approach taken by other public authorities to address-level property information.
But in the absence of any legal challenge or public statement from the ICO, which oversees compliance with data protection law in the UK, Land Registry’s view has been sufficient to underpin re-use of PPD. The open data release has been popular and useful.
However yesterday, in response to a Freedom of Information request from Madeline Bowles, Land Registry released a collection of correspondence with the ICO and others about the implications and impact of releasing Price Paid Data as open data. I’ve made PDFs from the original Word docs:
2004_01. ICO response 20 February 2004.pdf
2004_02. ICO response 2 April 2004.pdf
2004_03. 19 May 2004 17.44.pdf
2012_01. 15 March 2012 15.27.pdf
2012_02. 21 March 2012 15.58.pdf
2012_03. 21 March 2012 16.06.pdf
2012_04. 21 March 2012 16.14.pdf
2012_05. 21 March 2012 16.51.pdf
2012_06. 22 March 2012 15.24.pdf
2012_07. 22 March 2012 17.53.pdf
There are two groups of correspondence:
1. advice from ICO in advance of plans for commercial supply of Price Paid Data in 2004, and
2. further advice from ICO and discussion with Land Registry in the week immediately prior to the first open data release of PPD on 23 March 2012.
Most names are redacted but parties to the correspondence include Land Registry directors, legal advisors and someone at National Archives, as well as the ICO.
My interest in open data policy is admittedly wonkish, but I find some aspects of this correspondence frankly quite extraordinary. Here are the highlights:
In 2004 ICO really wasn’t sure
In the 2004 advice ICO does say of the Price Paid Data:
It would not appear that … the individual the focus of the information, rather it is the property to which the information relates.
But this is amidst a great deal of hedging (“This has involved much debate within the office …) and the letters raise some questions that don’t seem to have been resolved.
ICO presents its 2004 advice as a “preliminary” review and says it will write to Land Registry formally with “the considered opinion in due course”. As Land Registry has not disclosed anything further from ICO in 2004 I can only assume that this further opinion was never provided.
But Land Registry had what it wanted
We can see the legacy of this advice in Land Registry’s 2013 Privacy Impact Assessment. Constructively it seems as if Land Registry got what it wanted in 2004, i.e. sufficient coverage from ICO to go ahead with sale of the Price Paid Data.
In 2012 ICO is sure PPD is personal data
ICO position in 2012, in advance of open data release of PPD, is significantly more confident, and bad news for Land Registry.
Whilst we are supportive of Governments commitment to greater transparency, I reiterate our view that the data you plan to publish is in many cases likely to be personal data given the public availability of other data which would enable people to easily relate property price data to individuals who own or who are resident in the property. As the OGL does not cover the use of personal data, in our view, this is therefore not a suitable vehicle for the publication of property price information. As the licence has been developed by The National Archives you may wish to consult with them on whether they share your view on there being no personal data involved and whether the use of such a licence is appropriate. You may also wish to explore an alternative approach involving the use of a bespoke licence that could impose restrictions on using the information is a way that has adverse privacy consequences for individuals.
ICO acknowledges Land Registry’s contrary view (”we note that you do not agree with our view that the price paid data is personal data”) but says they will “still need to consider any complaints we receive from persons affected on their own merits.”
There was a bit of a flap
The remainder of the correspondence is internal emails between Land Registry directors, lawyers and the National Archives, culminating in a reply to ICO. Land Registry notes ICO’s view that PPD “may in many cases be likely to be personal data” but persists doggedly in its own view that “price paid data is property related rather than personal.”
Interesting side points
Many of the correspondents seem to think the fact that data is publicly available militates against it being personal. These days, with the rise of social media, most data practitioners take a more nuanced approach to that interaction.
The National Archives email includes the following line:
It certainly will not be possible to amend the OGL by tomorrow even if we were minded to do so.
Yes, it appears somebody actually wanted to change the Open Government Licence to resolve Land Registry’s problem.
What does all this mean for re-users of Price Paid Data?
As I see it, re-users of Price Paid Data have two options:
A. Take Land Registry’s position at face value and treat Price Paid Data, at the point of release, as non-personal data. That means you disagree with the position of ICO, the first line of regulatory authority on interpretation of UK data protection law. This is your decision. Neither Land Registry nor the Open Government Licence provide any indemnity to re-users for breaches of data protection.
B. Take ICO’s position at face value and treat Price Paid Data as personal data. That means you have no clear legal basis on which to re-use the data for any purpose. Land Registry has licensed re-use only under the Open Government Licence, and the OGL explicitly does not cover personal data.
B is the safest option, but that said I don’t want to overstate the risks attached to continued re-use of Price Paid Data under current conditions. So far the general public hasn’t taken exception to the availability of house price data. Provided there’s no legal challenge on data protection grounds, the status quo is likely to prevail. (But if there is a successful challenge, bear in mind that it could invalidate re-use of Price Paid Data retroactively. So I wouldn’t build a business model on the data.)
The big question
Why has neither Land Registry nor ICO made this dispute public previously? Land Registry’s PIA is a disingenuously one-sided treatment of the data protection issues. Both organisations have done open data users a disservice by leaving them to rely on a position that is at best open to debate.
Is there a better approach?
I advocated for open data release of Price Paid Data. I don’t think Land Registry’s characterisation of the data as non-personal is likely to stand up to legal challenge; however I would be pleased if it did.
Property information as personal data is a recurrent issue that has compromised the open data potential of a number of releases in addition to Price Paid Data, such as DCLG’s recent release EPC bulk data and Home Office’s release of fire incident data just yesterday.
In my view some of this data does technically fall within scope of the standard definition of personal data, but is not data that most people would necessarily recognise as personal. I think there is an argument for establishing, by legislative intervention if necessary, a special category of information about property (and perhaps also about other major private assets in which society has a legitimate interest) as outside the definition of personal data, even if it is notionally linkable to individuals.