Jersey’s Freedom of Information law was drafted in 2011 and a disclosure log was added to the States of Jersey’s website beginning in 2015, the same year the law came into effect. So far, so good. However it is easy to read the public data sets provided and to discover that hundreds of valid requests and responses are being omitted from the disclosure log. I decided to submit a new request to the Central Freedom of Information (FOI) Unit themselves, to ask why.
The response I received is incomplete and it leaves me with a few answers and some new questions. You can read PDFs containing my initial request, their response, and the (now properly redacted) attachments included with their response if you want to go straight to the source material.
The FOI Unit’s basic stance is that they are not legally required to disclose this public data, and that redacting information from requests and responses will take too much time. All responses are considered public, but because some contain data exempt from fully public disclosure and will require further redaction, it is important to consider both sides of the situation. On their web page they say all requests “…will then be anonymised and passed to the relevant business area that may hold the information you have asked for” so it is clear that all requests are already being anonymised by their office, and what we are really talking about is the time required to redact responses to valid requests.
Checking the documents and publicly disclosed figures, 2016 saw 736 valid requests, but only 612 appear in the Disclosure Log, a difference of 124. Of that remaining 124, a maximum of 37 are Planning Applications (this figure is important and is explained later) plus 22 others came in response to my request, leaving a remainder of 65. Most of those 65 missing FOI response documents for 2016 remain unexplained and are unaccounted for.
Checking the figures for 2015, 101 responses are missing from the Disclosure Log. An unknown number of these will likely be Planning Applications and the balance remain unexplained and are unaccounted for.
The Disclosure Log page says “Check that someone has not already asked the FOI question you plan to ask.” Because so many valid requests and responses are missing from the public log this is impossible.
Among the responses considered important enough to prioritise for public disclosure so far are requests for a zombie attack preparedness plan, some nutjob asking about dragon attacks, UFOs, and whether “Santa Clause”(sic) violates Jersey airspace, also nuclear attack plans, and the ever-important topic of “spit hoods”.
I don’t think all of the published responses are important and I question whether those that were published are worth devoting such limited resources to. I would prefer to see them all online, to me that is true freedom of information.
Jersey’s FOI law is a great step towards government transparency and the last two years have seen a majority of valid requests and responses published publicly online. Well done to the States’ website team and to the FOI Unit for making this data available. It is a great start but there are improvements that can still be made.
I sent the initial request 6th January:
My FOI request is about the partial list of responses published on the disclosure log on your website at this page: https://www.gov.je/Government/FreedomOfInformation/Pages/FOIDisclosureLog.aspx.
Can you please supply:
a) The criteria or official policy referred to when choosing which responses to post publicly on your website.
b) All previous responses to valid FOI requests which are not posted publicly on your website.
23rd January the FOI office rang me (from a private number) for clarification of my request, and to explain that as it was written, it would likely exceed the maximum 12.5 hours time limit to process it. I was told that the “vast majority of unpublished documents were Planning applications” and that redacting them would take too long. Fair point, so I asked for everything except the Planning applications. Confirming their intentions via email later the same day, they wrote “In compiling the information requested, we will exclude any responses to requests which relate to planning applications and in which either the request or the response also contain personal data.”
The bit at the end about any documents containing personal data was not disclosed during the phone call, they inserted that later after telling me that only Planning Applications would be omitted, which is what I agreed to. As they say, “words matter” and they deviated from the verbal agreement so I wrote back to clarify:
Thanks for ringing me earlier. As you have explained that the majority of unpublished FOI responses are Planning applications containing absolutely exempt personal data, and that the estimated time required to redact this data would exceed twelve and a half hours, I do not object to your suggestion that those documents be excluded from your response to my FOI request.
I do note that on the “How to make an FOI request” page on the States of Jersey website it says “Once information is issued to you in response to your FOI request, it is then considered to be publicly available.”(emphasis mine)
Since that data is not publicly available on the website, can you please confirm whether any of the following statements are accurate?
1. Some FOI responses contain exempt personal data upon release to the requesting party, and that personal data is about someone other than the requesting party.
2. Some FOI responses contain exempt personal data upon release to the requesting party, but only if that data is about the requesting party.
3. All FOI responses have all exempt personal data redacted before releasing them to any requesting party.
Apologies for being so specific here, but I am trying to make sure I have not overlooked anything obvious and I want to have a good understanding of the process. I’m guessing only number 2 is true and I hope you can confirm this.
At least their official written response says “As agreed with the requester on clarification, this number excludes the requests in relation to planning applications.” which confirms my point about what they told me on the telephone.
The FOI Unit’s Response
The response is effectively something like “We don’t have the time or staff resources to make this happen”. This will surely become fodder for local conspiracy bloggers, but in the absence of good data, people will speculate.
I won’t speak for them, instead I invite you to read their complete response for a fuller picture and to understand the FOI Unit’s perspective. They do make some good points but the public is once again asked to trust anonymous government employees to decide what is in the “public interest” and to withhold selected data sets while releasing important things concerning Santa and UFOs, etc… .
Then there is still a problem with the number of unpublished documents. Only 18 withheld requests are explained, of the 65 missing for 2016. The figures provided simply do not add up, they are not even close to being complete.
Many documents from valid FOI requests and responses remain unpublished. The explanations provided as to why this is true are incomplete, and the provided figures are impossible to justify with existing public data sets. The 101 missing 2015 documents were not even addressed in the response provided.
In speaking to the FOI Unit it is clear that producing a 100% complete Disclosure Log online is not currently achievable. That is disappointing to hear because the spirit of the FOI law is to disclose everything (excepting exempt data) and to leave it for the public to decide what is important.
Methodology & Data Sources
A total number of published responses is not made available, however there is a simple workaround available. If you visit the Disclosure Log page and click on a year in the “Filter by year” box in the right sidebar, a list of all published responses for that year is displayed on the page. Next, use your web browser’s native search functionality to match the exact phrase “Date published”, which returns 612 matches for 2016. That is how many responses were published, 612. An archived snapshot of this page is also available on archive.is.
Using the same technique for 2015 results, there are 552 matches (NB this excludes the 4th Jan 2016 result which also appears in the 2016 log - each result may only be counted one time). An archived snapshot of this page is also available on archive.is.
Using the States of Jersey published figures for all of 2016, 736 valid requests were made last year. An archived snapshot of this page is also available on archive.is.
Using the States of Jersey published figures for all of 2015, 691 requests were made that year. However it is not explicitly stated on the page whether these were all valid requests. My assumption is that they are, since they are presented alongside valid requests for 2016. An archived snapshot of this page is also available on archive.is.
Using the States of Jersey open data sets explorer we see the number of valid FOI requests for Planning applications in 2016 was 37. An archived snapshot of this page is also available on archive.is.
Using the original release of zipped PDFs that were emailed to me, I counted 22 files. I am unable to release these to the public due to improperly redacted personally identifying information. I don’t fancy an angry phone call from our Information Commissioner (already had one from the FOI Unit early this morning!) so I hope you will believe me when I say there were 22 files, which are now properly redacted and merged together into one multi-page PDF as the response attachment above.
A Final Note
Lastly, the PDF responses I received contain a number of FOI request rejections. That means they are not counted towards the total number of valid requests and so do not count towards the number of explained omissions. They aren’t responsive to my request since I only asked for responses to valid requests which were unpublished. Throughout this blog post I have been very careful to give the FOI Unit the benefit of the doubt in calculating the figures and the number of missing responses (65) are used even though this is the result of the most favourable possible interpretation of the figures, from their perspective.