OIA charging: Who charges for OIA requests in 2020?

Back in 2016, in response to a new OIA charging policy from the Reserve Bank and a push by then-Ombudsman Beverley Wakem for more agencies to charge, I conducted a survey of OIA charging practice. This confirmed the anecdata that in core government agencies charging for OIA requests was so rare as to be virtually non-existent (peripheral agencies and local government are a different story), and many agencies have implicit policies against charging. One of the purposes of the survey was to establish a baseline for further work, so we could see if this changed.

I've just completed another survey asking identical questions of each public service department seeking information on the number of requests they received, the number of times the demanded charges, the number of times they paid, and the amount collected. Responses were due back on Friday, and are collected here. The short version:

• Out of more than 9,752 requests made to 30 responding agencies, only 4 attracted charges, a rate of 0.04%. This is slightly lower than the 2016 result of 0.05% (excluding Customs, which has since stopped charging). Note "More than": one agency provided information on the number of charges, but not on the number of requests.
• One agency - MBIE - refused to provide information on the number of requests received, on the basis that different statistics would be released next month by the SSC. Appallingly, they took the full 20 working days to do so, which seems inconsistent with the OIA's "as soon as reasonably practicable" standard.
• Two agencies - MBIE and Treasury - claimed that they had no idea how often they charged, and that they would have to look through every OIA request to find out. Interestingly, both were able to tell me four years ago. Treasury accepted that this is something they should be tracking, and they are investigating doing so.
• One agency - the Ministry of Housing and Urban Development - did not respond at all, despite the request being sent to the main contact address listed on their website. A followup sent to their specific OIA address has produced a rapid promise to respond as quickly as possible, so it seems they have an internal administrative problem with passing on requests.
• 25% (versus 30% in 2016) of agencies responded more or less immediately, within 5 working days of the request.
• 40% (versus 30% in 2016) of agencies took over 15 working days to respond to a simple request for statistics which should be immediately available if they were tracking their requests properly. Both of these numbers are worse than last time, and raise questions about both these agency's commitment to the OIA's "as soon as reasonably practicable" standard, and about the performance trend of the government overall.

What can we conclude from this? The level of charging for OIA requests rightly remains practically non-existent, and Wakem's push to increase the rate of charging has failed. Agencies continue to recognise that charges are a barrier, and are rightly reluctant to impose them. That is a Good Thing, but what we don't know is the rate of requests declined for requiring "substantial collation and research" when a charge could instead have been requested. So in order to really assess the outcome, we need better, outcomes-based OIA statistics from SSC. Which would incidentally allow us to investigate a number of other issues - for example, the rate of extensions, and whether requests are being refused in order to meet timeliness targets.

Secondly, while its a sample of only 32 requests, the time histogram is troubling, and lends support to Mark Hanna's work suggesting that a number of agencies are treating the 20 working day maximum as a target, and ignoring the "as soon as reasonably practicable" standard. This is something that desperately needs to improve, and I'd like to see a work programme from the Ombudsman, SSC and Ministry of Justice to do so.

Educating schools about the OIA

Back in February, Christchurch Student Matthew Taylor began publishing the results of a series of OIA requests to schools about school formals and homophobia. The first result was that schools were unfamiliar with (and largely hostile to) their statutory duties under the Act, with over 150 not responding at all, and the majority of the rest demanding to know who he was and why he was asking. Now, the Ombudsman has published guidance on The OIA and school boards of trustees [PDF]. Among that guidance:

OIA requests can be made by any New Zealand citizens and permanent residents or anyone in New Zealand... requesters don’t have to be associated with the school. BOTs can ask a requester for information that is necessary to ascertain whether that person is eligible to make an OIA request. However, they should not impose unnecessary or unreasonable requirements.

(And even if they're not eligible, the response must still be "reasonable" - which generally means handling it under the OIA)

And:

A requester may give reasons for making an OIA request but there is no obligation on them to do so. BOTs can ask for this information if they think it would be helpful in reaching a decision, but they cannot insist that it be provided. BOTs should be careful not to suggest that the supply of reasons by the requester is a prerequisite for getting an answer to their OIA request. If a requester chooses not to state their reasons for making an OIA request the BOT must still fulfil its legal obligations under the OIA.

I think that's a pretty comprehensive spanking for the schools, and hopefully it will result in better OIA compliance in the future.

A reversal on secrecy?

For the past few years I've been waging war on secrecy clauses, submitting at select committee where clauses in legislation seem to over-ride the Official Information Act. One of the drivers of this was a 2014 decision by the Ombudsman (unpublished, but posted here), where they interpreted an exemption in one of the Climate Change Response Act's confidentiality clauses, which allowed information to be released "as provided under this Act or any other Act" as not actually allowing release under the OIA:

I am not persuaded that the OIA is an Act that provides for the disclosure of information in s 99(2)(a) of the Climate Change Response Act. The OIA confers a right to request official information and requires that such requests be processed in accordance with its provisions, but those provisions do not provide for the disclosure of information under the CCRA (or any other Act that imposes restrictions on the availability of official information). Instead, section 52(3)(b)(i) of the OIA provides that nothing in that Act derogates from any provision which is contained in any other Act which imposes a prohibition or restriction in relation to the availability of official information. Section 99 is such a section.

Accordingly, the OIA does not override the restrictions imposed by section 99 of the CCRA and it would be contrary to that section for the requested information to be made available to you. Consequently, section 18(c)(1) of the OIA provides a reason to refuse your request on that basis.

Since then such clauses have unfortunately become a regular feature of legislation, as business interests have lobbied for statutory secrecy to over-ride our democratic right to transparency.

There is a good argument that the Ombudsman's 2014 decision was incorrect, and that a BORA-consistent interpretation of the clause (as required by s6 BORA) would interpret it as giving effect to the right to freedom of expression, which includes the right to receive information and is internationally recognised as including a right of access to government information. And it seems that the Ombudsman has been persuaded. The new Offshore Renewable Energy Bill includes a similar secrecy clause to that in the Climate Change Response Act, with an exemption allowing release where disclosure is "required" (rather than merely "provided" by) other legislation. And in their submission, the Ombudsman says that this does not oust the OIA:

The Official Information Act 1982 (OIA) appears to be one such piece of ‘other legislation’ that may require commercially sensitive information or personal information to be released, on request under that Act.

They also highlight the constitutional nature of the OIA, and go on to say that:

The Ombudsman therefore generally is of the view that Parliament would only derogate from, or limit the application of, the OIA through clear and direct legislation to that effect.

Which is a similar approach to that taken by the courts on the BORA, or te Tiriti.

Taking the Ombudsman at their word suggests that many existing secrecy clauses - at least those prohibiting disclosure but with "required by" exemptions - may not in fact limit the Act, and certainly won't be interpreted that way by the Ombudsman should the matter come before them in a complaint. But of course there's only one sure way to find out...

The constitutional status of the OIA

The Ombudsmen Annual Report for the year ended 30 June 2011 [PDF] was released yesterday, and I've spent this morning reading it. OIA complaints reached an all-time high last year, though were down in many departments (the big exceptions being Corrections, Labour and MSD). John Key was the only Minister to attract more than 15 complaints. But the interesting part of any Ombudsman's annual report is the section on significant decisions. And this one has a very interesting one, about the constitutional status of the OIA. Someone had requested information about the citizenship status of an individual, and been refused on the basis that it would be contrary to an enactment - specifically the Citizenship Regulations 2002. The Ombudsman found that it would indeed be contrary to those regulations, but that it did not matter because the derogation was not explicit enough:

The Ombudsman accepted that release to the requester would be contrary to regulation 15. However, he did not accept that a regulation made after 1 July 1983 could have the effect of overriding the OIA unless the empowering provision under which that regulation was made expressly or impliedly provides for the presumption of availability of official information under the OIA to be abrogated. To accept otherwise does not do justice to the constitutional status of the OIA. There was nothing in the empowering provisions in the Citizenship Act 1977 that justified the setting aside of one’s obligations under the OIA. Accordingly, the Ombudsman did not accept that section 18(c)(i) of the OIA could apply.

In our view, Parliament must have made clear its intentions to override the OIA by way of primary or pre-existing secondary legislation in order to have that effect.

(Pre-1983 regulations are protected by the OIA's savings clause).

This is, I think, something new. But its also the way our constitution evolves. The courts have gradually taken this sort of view with both the Treaty of Waitangi and the BORA: that any derogation must be explicit and leave no room for doubt, because the constitutional (and in the case of the BORA, statutory) presumption is that Parliament would never derogate from these accidentally and therefore a consistent interpretation must be found. Now the Ombudsman is playing the same game on transparency. The result will be a more open government, and an even greater presumption against statutory secrecy. And that will be better for all of us.

The government's secret OIA plans

Back in September, when the government announced plans to increase proactive release of official information, we learned by accident that they were also considering another review of the OIA, and "intend[ed] to carry out targeted engagement to inform a decision on whether to progress a formal review". As someone interested in OIA reform, I was naturally curious about this, so I sent an OIA off to Justice Minister Andrew Little seeking information about the proposal. I finally got the response back on Friday, after a month-long extension for "consultation". Unfortunately, its not very informative.

You can read the released documents here. As is obvious, all interesting information about the proposal has been redacted. All their specific proposals for reform are secret, as is practically everyone they plan to consult in their "targeted engagement". People with specific expertise in the law? Secret. Bloggers and commentators? Also secret. They do list some media organisations, and the members of the OGP Expert Advisory Group, but everyone else is secret. Which is outrageous when you think about it. The OIA is quasi-constitutional legislation, something that belongs to (and affects) all of us. But rather than a full public consultation, they plan to privilege some voices over others, presenting their select secret proposals to a select secret group, then presenting the stovepiped results to us as a fait accompli. And they kept this entire process secret as well: they decided it all back in May, but never announced anything. The only reason we know about it at all is because of a passing reference in another document. Whether these are the actions of a government committed to transparency, accountability, and participation is left as an exercise for the reader.

As for the quality of their proposals, well, they're secret, so we can't tell. But what they do reveal isn't encouraging. For example, the Ministry of Justice's brief advice on whether the OIA should be extended to Parliament cherry-picks its examples to include only those that support the status quo, while ignoring the obvious counterexample: the UK Parliament is fully subject to its Freedom of Information Act, and this has brought about a huge improvement in accountability. Which doesn't provide much reason for confidence in the quality of their advice. And while they don't seem very keen on extending s48 of the OIA to cover proactive release, repeatedly highlighting the Law Commission's recommendation against extension, all their actual advice is secret, so we can't tell whether its robust or flawed. Which given the huge potential for abuse in the proposal - it would basicly give Paula Bennett total impunity to dox people at will with their benefit, medical, police and tax records - is something we need to know.

We deserve better than this. Its not just politicians, journalists and trouble-making bloggers who use the OIA, but all of us. Steven Price's 2005 study of the OIA contained an extensive list of examples of how ordinary citizens use the Act, and summed it up as "the stuff of democracy". According to the Ombudsman's 2017-18 annual report, individuals made three times as many OIA complaints as journalists, and its 5.5 times as many when you look at the LGOIMA. In short, it's our Act, not theirs. And any non-trivial changes to it require publicly consulting all of us, not just a select group of chosen insiders.

Update (31/1/19): Andrew Little has reconsidered his decision and released the options under consideration. More information here.

Pissing on the OIA

Earlier in the year, David Hines released the findings of his remarkable OIA project on religious education in state primary schools. He sent requests to all 1800 state primary schools. What's appalling is that over 1000 of them were "reluctant" to answer and had to be persuaded to by the Ombudsman. In February, 260 were defying the Ombudsman's demand to respond to Hines' request. And now they've gotten away with it:

This Office received a complaint about the lack of a response by over 1,000 schools to a request for information (in the form of a religious education survey). We took an active role in supporting the Boards of Trustees to better understand their obligations under the OIA and, as a result, most subsequently provided the information requested. However, over 100 Boards of Trustees still failed to provide a response to the requester despite this being in breach of their obligations under the OIA.

The requester chose not to continue pursuing his complaint against the relatively few Boards of Trustees that remain in breach of their OIA obligations.

So, they ignored the law, and then they just waited it out until the requester got bored and decided to move on. If you need an example for why we need criminal penalties for breaches of the OIA, this is it. We can jail people for contempt of court; the Ombudsman should have similar penalties available for errant public officials.

I'm trying to get a list of the schools so they can at the least be named and shamed. These schools are failing both in their legal duty under the OIA, and also in their responsibilities as educational institutions to mould our future citizens. The message they're sending their students is crystal clear: the law does not matter. The sheer hypocrisy of that position (created by law, funded by law, and purporting to impose rules on others) should be plain to all.

And OTOH, is this really surprising? The Ombudsman is now so underfunded that it is reportedly taking six months for them to even give OIA complaints a number, and a complaint can take two years to resolve. Under these conditions, it is no wonder people give up waiting for justice - and no wonder agencies have started ignoring the law.

If we want the OIA to work, we need to properly fund enforcement. But Ministers have no interest in that, any more than Turkeys have an interest in christmas dinner. And so the OIA, a core part of our democracy, dies a slow death from neglect.

Using austerity to attack democracy II

Back in March, Newsroom's Sam Sachdeva reported that the regime was planning to use high costs as an excuse to limit access to the Official Information Act, and had commissioned consultants to build them a case. Obviously, I was quite curious about this, so I fired off a series of OIA requests to government agencies. Some were about costs, and revealed (unsurprisingly) that no-one (at least, no-one in the sample of core government agencies I checked) had any idea how much the OIA cost them, and that no-one was counting. Some, aimed at understanding recent large increases, were about the types of requests received, and when they had started being counted in statistics, and they were quite illuminating. And of course there was a request for the policy advice underlying the regime's plans. A request for Te Kawa Mataaho's advice got bounced to Ministry of Justice - saving me the effort of asking them directly - where it was delayed, and then delayed, and then delayed again. They finally responded yesterday, having illegally delayed the response until a related proactive release was signed off.

The response letter, with links to the released documents, is here. There's a lot of emails (some of which are significant), a couple of contracts, and notes from a meeting with the Ombudsman which suggests the regime's intent is to use this as an excuse to declare requesters "vexatious" - something both the Law Commission and the Ombudsman have opposed in the past. The Ombudsman is clear, both there and elsewhere, that one of the drivers of costs is increasingly convoluted agency review and sign-off procedures, where every response is scrutinised by multiple layers of management for arse-covering and "no surprises" reasons. This frequently results in delays, and it has been a frequent topic of the Ombudsman's practice reviews.

As for the costs themselves, they're in the proactively released Tregaskis Brown report and accompanying briefing. You hire consultants to deliver the answers you want, and TBL has delivered in spades, with a headline cost estimate of

$183.6 million for the 2024/25 financial year, within a possible range of $175 million – $250 million depending on estimated complexity of OIA requests.

They've calculated this by (roughly) taking the number of requests, estimating the proportion per agency which are complex (expensive) vs routine (cheap), and multiplied by the relevant cost per request, based on Australian data. They correctly highlight that 77% of requests come from 6 agencies (Police, Natural Hazards Commission (EQC), Department of Corrections, NZ Defence Force, Fire and Emergency NZ, and NZ Customs Service), and (sensibly) recommend further work to understand both actual costs, and what is driving them, as well as strengthening proactive release. Those bits are fine, but the cost estimate is absolutely absurd. How? The accompanying A3 (p10 of the proactive release) estimates the police's annual OIA costs at $50 million. But the police are actually one of the few agencies we have good OIA cost data for! A November 2025 request made on FYI, the public OIA request site, included both the numbers of requests processed, and the staff numbers and costs for both the police Ministerial Services OIA group, and the Information Requests Service Group (IRSG), for exactly the time period TBL is looking at. Ministerial services processed 1014 requests in the second half of 2024, and 832 in the first half of 2025, for a total of 1846 in 2024-25. IRSG processed 28921 and 25280 respectively, for a total of 54201. (Yes, there's a discrepancy between these numbers and those published by TKM. The reason for that is speed cameras and media requests, which are handled by other groups).

As for costs,

Police advises that there are 13 people in Ministerial Services who process OIA requests. As at 1 November 2025, the annual total for their salaries is $1,547,991.

There are 53 people who work in the Information Request Service Group and the annual total for their salaries is $4,489,296.

This gives a cost-per-request for police ministerial services (which handles the most complex requests) of $838.57, and for IRSG as $82.83. While there's no cost-estimate for speed cameras and media requests, both are likely to be at the lower end (the former because they are routine, the latter because anything non-routine gets kicked into the formal OIA process and handled by Ministerial Services). These are obviously far lower than TBL's estimates of $3530 for complex requests and $353 for routine ones. Even allowing for overheads (which TBL estimates at 66%), TBL's costs are inflated by a factor of 2.5.

You would hope that TBL's followup work will reveal that. In the interim, though, Goldsmith has got what he paid for: a big, scary number for costs, which he can decry as "waste", plus the inevitable line-go-up graph, showing those costs will increase into the future. Which he will probably consider to be a case for action.

The proactively-released briefing notes that the OIA is of constitutional significance, and that any change will attract significant interest. It recommends 10-12 weeks of public consultation on any proposal. That's clearly not happening on the original proposed timeline of "before July", so it may have been kicked back until after the election. Alternatively, Goldsmith being Goldsmith, he may just not bother with proper process. This regime has established a terrible reputation for ignoring advice and enacting radical, anti-democratic, even constitutional vibe-based policy under urgency. Sadly, we can't rule out them acting as they have in the past, and wrecking the OIA in the same manner. The only way to stop them is to throw them out of office as quickly as possible.

The Commerce Commission's weak case for secrecy

Back in September, the regime announced plans to give new powers to the Commerce Commission. But the announcement also included this:

We have also heard in your submissions that businesses and individuals are increasingly reluctant to share information with the Commission because of fears confidential information could be released under the Official Information Act, potentially leading to retaliation or misuse of confidential information by competitors. This is undermining the Commission’s ability to collect evidence and receive useful information, particularly in investigations and merger clearances.

The regime's solution was of course more secrecy, with a 10-year blanket exemption from the OIA for "confidential" information provided to the Commission, and greater power for the Commission to issue temporary exemption orders. I was curious about the justification for this, so I asked the Commission whether they in fact had any evidence supporting it: were they aware of any OIA release from them actually causing the harms the Minister had alleged, and did they have any evidence their existing secrecy powers were inadequate? In both cases, the answer was "no":

Regarding the first two bullets of your request, the Commission is not aware of any specific instances where information we have released under the OIA has caused harm to the business who provided the information to us.

The Commission is also not aware of any documents containing specific evidence that section 100 of the Commerce Act is inadequate.

What about wider advice on the OIA? Here the Commission said they had information, then refused to provide it for a further two months as they were (illegally) "consulting MBIE and the Minister’s Office prior to making our decision on the potential release of this material". But they finally provided the response yesterday, and a folder full of documents. There are a few interesting things in here, including that the Commission has apparently been running its own private "special advocate"-style system for merger cases, where lawyers are given access to evidence but forbidden from discussing it with or disclosing it to their clients - similar to the system used in "national security" cases here and overseas, with all the unfairness and professional issues that entails, only without any statutory authorisation. But on the actual case for secrecy, its largely fear, uncertainty, and doubt. TL;DR businesses are afraid they will be harmed by the release of "confidential" or commercially sensitive information. There's also fear over the public interest over-ride, and the inability of the Commission to give categorical assurances of total secrecy. Both show that businesses do not understand the law (which is to be expected), but that the Commission seriously entertains this shows that they don't either (possibly due to corporate culture capture). The fact is that there is a clear and obvious case for withholding confidential evidence under s9(2)(ba)(i) (in that it is clearly in the public interest that people are able to give evidence to the Commission, so if release would inhibit the giving of such evidence in future, s9(2)(ba)(i) applies), and while this is subject to the public interest test, the reality is that in practice such information is almost never released, because the usual public interest factors of accountability, transparency, and participation simply don't apply to information provided by third parties about themselves.

(There is the issue of the accountability of the Commission for its decisions, which means they must release the evidence which justifies them, but they should be doing that publicly anyway, so that's not an OIA issue, but a basic one of administrative law...)

However, there is one significant issue: big companies intimidating smaller ones from giving evidence against them:

In cases involving an applicant with alleged market power, dominance, or some other form of power or leverage over market participants, those market participants may be particularly concerned by the prospect of any information provided to us being provided to the applicant. This is of particular concern to us, as cases of this nature generally merit scrutiny.

Which sounds reasonable at first glance. But it isn't specifically an OIA problem - because, as the Commission admits, it is required to provide such information to applicants for reasons of natural justice. So the applicants are going to find out whether a request is made or not, and all attacking the OIA does is hide information from other people.

The obvious move here is not to undermine the OIA, but to target the actual problem of retaliation and victimisation, just as we do for whistleblowers. And the government announcement included that, so there's no need for secrecy at all.

The release also includes a summary of public submissions to a consultation by MBIE, which gives a good overview of their consultees' views on "protecting confidential information". Its worth noting that a broad OIA exemption was not one of the options canvassed in that consultation, so the Commission is going well beyond what was floated. Its also shocking that any government agency would fail to recognise the constitutional nature of the OIA, and that their response to it causing them minor irritations is to try and exempt themselves from a fundamental part of our constitution. But again, this is likely a matter of capture by corporate culture. We know that local and international business are fundamentally hostile to democracy and transparency; its utterly shocking that the body we have established to police them has been so captured by them as to share that hostility. At the end of the day, the Commerce Commission is a public body. That means it must respect democratic norms - including the OIA.

Is OIA handling under-resourced?

One of the most common excuses for OIA failures is under-resourcing. Government agencies sometimes claim that don't have the staff or the time to properly obey the law. Which might have worked as an excuse briefly when the OIA was new, but its been the law now for 40 years, and you'd expect them to have sorted their shit out long ago. More importantly, ministers and chief executives have a legal duty of stewardship, and are required to ensure that their agencies can properly perform their statutory functions. If they do not have the resources to perform those functions, then it is the responsibility of its chief executive to ask for them, and of the minister to ensure they get what they need.

But do they actually do this? I asked 34 government agencies - the 32 public service agencies scheduled in the Public Service Act, plus NZDF and the police - whether they had proposed budget initiatives seeking additional resources for OIA handling, information and records management, or ministerials for any of budgets 2018 - 2021 (OIA handling being what I was after, ministerials because OIAs are often done by an agency's ministerial team, and information and records management because these are important supporting technologies which enable requests to be processed). I did not ask internal transfers because I did not think of it at the time, but some agencies were either helpful or realised it would make them look good and provided that information anyway. I did not ask about budget 2022 because that part of the request would have been refused.

I received the final response yesterday, 41 working days after the requests were lodged. A summary of the responses can be viewed here. Of the 34 agencies:

• 29 had not made any such budget bids and did not volunteer any other information (these are classed as "informative refusals" on the spreadsheet, and their specific comments are included as a note. The responses were largely template, and MBIE's is typical).
• Four agencies - the Department of Conservation, Ministry of Pacific Peoples, Ministry of Health, and the Police - had allocated additional resources from baseline funding or by a formal internal transfer, but had not asked for additional resources via the budget process. In DoC's case, they'd actually done it twice within the relevant time period, and provided a full business case for each as well as an interesting look at how various other agencies resource OIA handling. It is possible some other agencies also provided resources this way; if they didn't take the opportunity to tell me about it, that's on them.
• Only one agency raised a formal budget bid: Oranga Tamariki. And it was really about privacy requests, rather than the OIA.

So there you have it: fewer than one in eight agencies allocated additional resources to OIA handling, and only one raised a formal budget bid remotely connected to it (and that was really about privacy requests). The conclusion is that agency chief executives feel they are properly resourcing OIA handling - or perhaps that requests to ministers for additional resources would not be welcome. But regardless of the reason, the fact that they have not asked for more undercuts any claim to be "under-resourced". This myth is busted.

Judith Collins' two-tier OIA service

Back in August, we learned that sewerblogger Cameron Slater was receiving extraordinary OIA service from then-Minister of Justice Judith Collins, in one case receiving a response to a request within 37 minutes. But it wasn't just extraordinary for its speed; from OIA records OIA'd via FYI, it appears that Slater's OIA requests were not even logged.

According to 3News, Slater's afterhours OIA for a letter the Minister had only just received was handled on December 21, 2012. Here's the relevant section of the logs:


Slater's OIA request doesn't appear. Which is highly unusual and suggests it was handled outside the normal process.

Slater's 37-minute OIA doesn't appear either, but two similar requests do:


(Thanks to @LostArcNZ for the excerpts and the legwork)

Neither matches either the final request - for correspondence since August 2012 - or the date of response: 12 February 2013. Its possible that the first request could be the relevant one, but then you have to ask why Collins' staff put the wrong response date in.

What's going on here? It's pretty clear Collins was operating a two-tier OIA service: one for the public (official, logged, slow), and one for her pet sewerblogger (unlogged, lightning fast, with special tips on what to request and likely distorted release decisions). That's a gross abuse of power as a Minister, and to the extent that she politically profited from it, a corrupt use of official information.

A mini-Official Secrets Act for the Reserve Bank

The Finance and Expenditure Committee has reported back on the Reserve Bank of New Zealand Bill. Part of the bill is a new confidentiality clause, restricting publication or disclosure of information obtained using the bank's information-gathering powers. The first-reading version was actually good, with explicit recognition of the OIA. So naturally, the Labour-dominated committee reversed that, to specifically limit the application of the OIA:

The Bank may make information or data to which this section applies available under the Official Information Act 1982 only if 1 or more grounds under subsection (2) apply.

The scary thing is that this is actually an improvement on the current law, which excludes the OIA entirely.

How did this happen? The committee is "concerned that confidential information could be released under [the OIA]". But delving into the advice, it appears that the Reserve Bank in their initial briefing (p 21) thought the initial clause thought that it already did, saying "The confidentiality provision has the effect of limiting the availability of information under the OIA" (despite the clause saying the opposite). But their later departmental report recommends a "clarification" as:

The policy intent behind this provision is that information would only be available under the Official Information Act 1982 (OIA) if it could otherwise be released under this confidentiality provision. This ensures that information that is compulsorily acquired from financial institutions is appropriately protected. The provisions in the Bill are based on section 135 of the Insurance (Prudential Supervision) Act 2010, which had the same policy intent. Engagement with the Office of the Ombudsman suggests that amendments to this drafting would be desirable to more clearly give effect to this policy intent.

Both the bank and the committee believe that ousting the OIA is necessary to protect this information. But why does it need to be protected? The initial briefing says the RBNZ asks banks about their credit card rates or the amount they have loaned out to do things like help set monetary policy. Obviously, there are interests around bank regulation, commercial prejudice, confidentiality, and potential economic damage if this information is released. But the OIA already has provisions protecting those interests, in sections 6(e), 9(2)(b), 9(2)(ba), and 9(2)(d). The problem here is that neither the RBNZ or the government (or apparently the Ombudsman) trust the OIA to do its job. So instead, they want to give the bank its own mini-Official Secrets Act instead.

This control-freakery is unnecessary and undemocratic. But isn't it so very, very Labour?

Why we need an OIA review

Stuff has relaunched its "redacted" series on freedom of information, starting today with an examination of OIA delays, and a look at the government's broken promise of an OIA review. The latter has raised some doubts about whether a review is really necessary, including from Ombudsman Peter Boshier (who thinks its all a matter of bad public servants and that everything would be fine if they were just good chaps), and I agree there are real questions about whether the government can be trusted to meddle in transparency legislation (especially given the attitudes displayed by Ministry of Justice, the agency responsible for the Act). But for those doubting whether we actually need one, there is a very simple answer: the law is now 40 years old.

The OIA is, bluntly, the product of another era. And not the grim, pinched, haunted-by-a-piggy-cackle one we think. The attitudes which shaped it are even older. Alan Danks, the man whose name is on the report which gave us the OIA, was born in 1914. While his report Towards Open Government is a highmark of a liberal trend which had been opening New Zealand up since the 1960's, its also clearly shaped by the authoritarian, monarchist, deferential, anti-democratic attitudes of the society he grew up in, and which he helped undermine. A society where government was top-down, not bottom-up, distant and accountable only at elections (which effectively produced a triennial dictatorship), where foreign affairs and defence were "high matters of state" for kings and their anointed ministers, to be kept out of the mucky hands of the peasantry. Most importantly, a society where all government information was presumed to be secret, and its disclosure criminal.

The OIA (and Danks) changed that, and helped make us into a better, more open, more democratic society. Today, the attitudes underlying many of its withholding grounds seem highly questionable, if not downright suspect. Modern people approaching the same problem are likely to find a different balance between secrecy and transparency than Danks did - and one more in favour of transparency, because it has in many ways become a default.

But its not just about social attitudes. We've had massive constitutional change in that 40 years, most obviously in the form of the BORA and MMP. And the BORA is a particular issue, because it explicitly affirms the right to receive information. Meaning every OIA withholding ground prima facie violates it, and needs to be re-examined to see if it really is a reasonable limit which can be demonstrably justified in a free and democratic society. Its been the law for 30 years now, but that exercise has never taken place. A review would let us do that.

And of course there's the fact that the rest of the world hasn't stood still. The OIA was one of the first right to information laws passed. Now there's 40 years progress overseas with a number of valuable lessons. Other countries make their parliaments, their courts, and even private bodies receiving substantial public funds subject to transparency laws. They have better complaints and enforcement mechanisms. They have legal requirements for proactive publication. We are decidedly behind the times.

None of this of course means we can trust the government to meddle with the OIA. Labour and the Ministry of Justice (the agency responsible for overseeing the Act) have shown their hostility to its values, and National isn't any better. Any review must be kept out of the hands of politicians and officials to ensure it is not captured and used as a vehicle for pushing greater secrecy. It must also be public and transparent, allowing for genuine consultation. That means - Labour - consultation when a policy is being shaped, not afterwards, as a rubberstamp (and it certainly does not mean "select committee submissions". They're a final fix-up of an end-product, not actual consultation). Which suggests that maybe we should just cut the government out of the process entirely, have NGOs and civil society organisations interested in transparency run their own review, use it to draft a bill or series of amendments, and then campaign for its passage.

Parliament says "no" to transparency

The Parliament Bill Committee has reported back on the Parliament Bill. As usual, they recommend no substantive changes, all decisions having been made in advance and in secret before the bill was introduced - but there are some minor tweaks around oversight of the new parliamentary security powers, which will likely be shown to be inadequate within a year or two. As for my major theme - extending the OIA to Parliament - the committee basically said "fuck off":

We note that previous reviews, including by the Law Commission, have considered a possible extension of the OIA to cover Parliament. We also acknowledge the calls from submitters to extend the OIA to cover more parliamentary information, a view that some members of the committee generally support and would like to see progressed.

However, the bill as introduced does not amend the OIA, and for reasons of scope we cannot recommend substantive amendments to that Act. A full policy process would be required to ensure any proposal would not adversely affect the political, policy, or constituency work of members and political parties, nor the ability of the House to maintain control over its own proceedings. Moreover, a reliance on the definition of “proceedings in Parliament” from section 10 of the Parliamentary Privilege Act may not be suitable in the context of the OIA.

Firstly, hiding behind scope is bullshit - it is entirely normal for select committees to amend the schedules of the Ombudsmen's Act or OIA to add agencies which have been excluded. As for the need for a full policy process, this is basically an admission that they haven't done one - that despite recommendations stretching back to the Danks Committee in 1980, they didn't bother to consider the issue when developing the bill. Which is a hell of a failure in the policy development process - but I guess what you get when you develop major legislation in secret and without any public consultation.

There is more about this failure in the bill's departmental report (p22), where after reiterating all the whining about how they couldn't do it in the past, and doing a bit of scaremongering about what they've been asked about and therefore what they might have to release, they basically say "we do not administer the OIA, and any such policy project should be undertaken in conjunction with the Ministry of Justice". Well, they don't administer the Privacy Act either, but they were perfectly capable of consulting when they planned to extend it to cover information held by Parliamentary Security. So it does basically seem to be a prolonged case of "don't wanna" from an institution which has always felt itself to be above the laws which apply to others.

The committee does talk about the Protocol for the release of information from the parliamentary information, communication and security systems as a substitute for the OIA regime. Except when you read it, most of it is about secrecy and MP's veto power over the release of any information relating to themselves, and the bits covering general requests and information about parliamentary administration are either very vague, or entirely at the discretion of the Speaker. Still, there are obvious things to ask about, and we can see if the transparency they are claiming actually exists, or whether it just exists in theory as a way of defending against real, enforceable transparency.

I should note that one are where there might be more transparency is MP's expenses, where the Speaker will effectively get a regulation-making power to decide what will be reported publicly. But against that, the Speaker is an MP, with huge conflicts of interest around the making of such regulations (both because they have expenses themselves, and they need to maintain relationships with their caucus and other parties). Again, we can wait and see if that actually amounts to anything more than empty promises.

Meanwhile, as for those members of the committee who support bringing parliament under the OIA, I suggest speaking up about it, and putting a member's bill in the ballot enacting the Law Commission's proposed changes to start the process. I'm more than happy to draft it if they need help.

Rules for effective OIA requests

After seeing one too many terrible requests over FYI, I've tried to put together some basic rules on how to make an effective OIA request. These are really aimed at beginners, and intended to help them avoid unnecessary frustration by avoiding common pitfalls (because there's more than enough frustration in the OIA process, and we can at least avoid inflicting it on ourselves).

1. Don't rant: I've seen far too many requests introduce their subject with an angry tirade, filled with hostility and accusations. Don't do this. For a start, its being an arsehole, and that's never a good approach when you want something from someone. While "requester was an arsehole" is not a valid withholding ground under the OIA, it certainly primes the people processing your request to think of it as vexatious from the start, and it is not going to encourage them to be extra-helpful in the way you want them to be. Secondly, its a distraction. A frothing rant may feel good when you type it, but it distracts from the focus of your request, and makes it harder to work out what exactly you are asking for. And you will likely get a worse answer as a result.

   What should you do? The OIA is a process for requesting information, so request information. If you are asking for a specific document, don't waste space on a preamble, just ask for it. If your request is less specific, then you may need to provide context, but keep it to a minimum. Lots of my requests are sparked by media stories, so I usually go for an intro like "Today Stuff reported that X (ref to headline so they can see what I'm talking about). I would like to request the following information under the OIA..." - and that's enough. You don't need to write a huge introduction, let alone an angry rant.

2. Know exactly what you are looking for: This is easier said than done, and not helpful advice to beginners, but pretty obviously if you know how government works and what sort of information it generates and where, you can get more value out of the OIA and be more effective.

   What if you don't? I didn't when I started out, so I made requests asking for "all advice and communications" on a particular policy. That's a useful phrase, but in some cases it may turn up a bit much, and the agency may ask you for a specific timeframe or whether you want emails. If you're not sure exactly what you're after, its better to narrow the request rather than risk a refusal for "substantial collation and research". After all, you can always ask for the other stuff later. And as you build up specialist knowledge in your area of interest, you'll get more of an idea of what might be there to find.

   Whatever you are requesting, be as clear as possible about it, since this will avoid mutual frustration over misinterpretation later.

3. Ask the right agency: If you ask the wrong agency, they have a duty to transfer the request, which is another 10 working days and some unnecessary frustration. You can avoid that by asking the right agency in the first place. Again, knowing how government works and which agency does what helps here. If you don't, who to ask should be apparent from the circumstances that provoked the request (a media story will usually suggest who to ask, a consultation process will be run by some particular department, and if you've had unsatisfactory dealings with an agency then they will hold the information about their policies and procedures). If you're still not sure, there is a thing called the Directory of Official Information, which lists all government agencies, what they do, and what types of information they hold.

4. Know what you can get: The OIA includes numerous withholding grounds, and some of your request may be redacted or refused (you can challenge such refusals via the Ombudsman, but that's really beyond the scope of this post; if you want to learn to be effective at that, start reading the Ombudsman's guidelines)

   Requests for the following information will almost always be refused, and so it probably isn't worth asking for unless you specifically plan to challenge a refusal:

   • "national security" stuff
   • specific communications with other countries
   • details of active criminal investigations / court cases
   • upcoming budget docs
   • legal advice

   Other things you shouldn't bother asking for: court records, and stuff clearly done by politicians in a "political" (party leader / MP) or personal capacity rather than in their capacity as a Minister (e.g. coalition negotiations, where they went on holiday, what they had for lunch). Neither of these are "official information" in terms of the Act, and you can't get them via the OIA.

A good OIA story

I've had a lot of bad OIA experiences with the police. Quite apart from the vaguely sinister phone call I got from them once in response, there's the litany of requests "lost", late, or answered with glib bullshit. The police regularly top the list for OIA complaints, attracting over a hundred in 2008 - 2009. Institutionally, they seem to have an attitude problem, an entrenched culture of secrecy which holds that they should not be accountable in any way to the public and that we have no right to know what they are doing.

So its good to see that some of them are doing it right. Over the weekend, I acquired the 2010 OIA tracking data for the Auckland Police District. In the six months to June 31 2010 this district had handled 2279 requests. Six of these had poor data (e.g. typos in dates) so I don't want to guess at them. Of the remaining 2273, only three were late. The median response time was one working day. Its certainly a contrast with Police National headquarters, who in my experience are always late and can't find their arse with both hands.

The real story here is really that most OIA requests to police are routine. While I asked for the information on the subject of requests to be excluded to protect privacy, I understand that these requests are mostly people getting their file for their court date or insurance companies finding out if that burglary happened so they can pay out. While these have always been part of police business, they're technically OIA requests (every time you ask the government anything its technically an OIA request, and they have to help you with it; you don't have to say the magic words "Official Information Act"). And for this sort of stuff, the OIA seems to be working very well indeed.

The OIA is not a petition mechanism

Last year, we saw an organised campaign using FYI [down, but will be up again] to lodge OIA requests with Tim Groser demanding the release of the draft text of the TPPA. The aim was to show support for release, effectively using the OIA as a petition mechanism. All it resulted in was 50 form-letter refusals.

And now there's another similar campaign, this one over EQC. Unlike the previous one, which asked people to lodge their own requests, this one asks people to "add their electronic signature in support of an Official Information Act (OIA) Request". The implication is that this will lend the request more weight. It won't. Instead, its a pointless exercise which mistakes the OIA for a petition process. But that's not what the OIA is about. The purpose of the Act isn't to help people voice their grievances about the government, but "to make official information more freely available". Whether information is released is (supposed to be) decided on the merits and in accordance with the principle of availability, not on how many people want it. So unless you're the first person to request something, or you have another angle which might excavate different information, "adding your signature" to an OIA request or submitting duplicate requests has no effect whatsoever - and to the extent it results in form responses, it actually undermines the latter.

As for the request itself, it falls into all the usual pitfalls: a political preamble which invites the responding agency to treat it as correspondence rather than a request; a question-and-answer format; trying to get all the answers at once. These are pretty much guaranteed to result in an unhappy OIA experience:

An ordinary practice

Shock! Horror! Someone is using false names to make OIA reqeusts!

False identities were used by a right-wing lobby group to make Official Information Act requests of a government agency.

The Herald uncovered evidence showing people seeking information for the NZ Taxpayers' Union did not actually exist - and that numerous email addresses from its purported members were actually directly linked to its head office.

One single Taxpayers' Union email address was linked to nine ghost people who filed OIAs seeking information, including details later used by the lobby group to seek publicity.

The Herald approached the Taxpayers' Union for comment. It refused to talk for two days then issued a statement admitting it used bogus identities to make OIA requests.

Of course they do. So does every serious requester. In fact, I would be extremely surprised if the Herald's David Fisher, who produced this article, hadn't done it himself. As for why, the reason is simple: because there is a well-founded perception that agencies treat requesters differently based on who they are and the reason they believe they are requesting information. A request from a journalist or advocacy group may be delayed, transferred, or see as much information withheld as possible, while a request for exactly the same information from a random member of the public will see it speedily released. So regular requesters often pretend to be other people to get the information they are seeking.

Its a practice as old as the OIA itself. I recall reading an ancient article about the toxic days of the Fourth Labour Government, where people in one Minister's office were having to make pseudonymous OIA requests to find out what other Ministers were doing. Nowdays, with email and FYI, its trivially easy. Make a throwaway account, file OIA request, wait 20 working days. Its only problematic if the agency is one which uses eligibility requirements as a barrier, or if you want to complain to the Ombudsman later.

Is it legal? The Law Commission noted in its 2012 review of the Act that there's no requirement for requesters to provide their real name, and agencies would have no way of knowing if they did anyway. It therefore made no recommendation on the matter. The Ombudsman is on record (in their submission to that review (Q45)) that there should be no requirement for requesters to state their real name. So its certainly not seen as a problem by those responsible for upholding the OIA regime.

In other words, rather than some weird and underhand practice, its a common OIA tactic, and likely used by the journalists themselves. Rather than investigating it, maybe the Herald should be looking at why it is necessary, and campaigning for a better freedom of information regime, rather than persecuting those struggling to make it work.

(Meanwhile, the Herald's investigation seems pretty dubious itself. They obtained the email addresses of requesters, though its not clear how, and then attempted to hack the accounts to learn their password reset emails. FYI is on record as saying they did not provide the addresses, and I think there's interesting questions to be asked about how the Herald did it...)

Edit (4/10/18): David Fisher has got in contact to assure me that he has never made a pseudonymous OIA request. I guess I get to be surprised then.

The "most open and transparent government ever" again

One of the good transparency moves over the last decade has been an increasing shift to the government releasing information proactively. We now assume as a matter of course that the cabinet papers and briefings underlying every significant government decision will end up on an agency website somewhere. These documents usually contain redactions, and usually refer to a justification under the OIA for doing so, but they're available, without anyone having to request them, and that's good.

But this isn't just driven by a desire for transparency: proactive releases aren't actually covered by the OIA, so those redactions don't actually have to be lawful under the Act. And this is used by the government to keep information secret and gain tighter message control. Newsroom's Marc Daalder has a perfect example today of a Minister caught in the act of censoring government information to keep it out of the public eye:

Climate Change Minister James Shaw redacted parts of a Cabinet paper he released last year, even after officials advised him not to.

The paper predominately dealt with a decision on price controls in the Emissions Trading Scheme (ETS), which Shaw announced at the same time he released the document. But it also contained reference to an upcoming review of the ETS which Cabinet had yet to make final decisions on.

According to a separate briefing released to Newsroom under the Official Information Act (OIA), officials from the Ministry for the Environment told Shaw there were no valid grounds under the OIA to redact the reference to the ETS review. Releasing the paper would therefore effectively announce the review. Alternatively, they advised, Shaw could choose not to release the paper.

Instead, Shaw went a third way - releasing the Cabinet paper but redacting the mentions of the ETS review against official advice.

If the paper had been requested under the OIA, Shaw would not have been able to hide this information, and his decision to do so would have been reviewable by the Ombudsman. But because he proactively released it, he could do what he liked, and hide whatever he wanted to keep the public in the dark.

This is bad, but the real problem is that the fact that a document "will soon" be proactively released is a reason to refuse a request for it. But then, that eventual proactive release can contain redactions which are not lawful under the OIA. Shaw has been particularly assiduous about refusing things on this basis, but what we've seen here is that we can have no faith that his eventual redactions will be lawful. And he's not the only one - a recent case note from the Ombudsman found the Minister of Health was also playing this game, and as a result the Ombudsman will be taking a closer look at s18(d) refusals generally. Because you can see how this undermines trust in the entire OIA system.

As for how to fix it, the short-term solution is to request unredacted copies of every proactive release. I've found it useful when doing this to ask that the request be handled urgently, to note explicitly that saying that they stand by their redactions should not take more than a few days, and to explicitly threaten an "undue delay" complaint if they piss around (setting a clear path-of-least-work incentive for lawful behaviour). Alternatively, Daalder has shown that requesting the briefings on specific proactive releases can be fruitful and informative.

The long-term solutions are twofold: the OIA needs a formal framework for proactive release (the UK FOIA's model of publication schemes seems to be a good one to steal), and proactive release needs to be reviewable by the Ombudsman. This needs to include not just redactions, but also non-compliance with a publication scheme - failing to release something on time as expected. This solution would take years to put in place, but in the meantime it seems easy enough to add reviewing redactions in proactive releases to the grounds of complaint to the Ombudsman, to ensure they are lawful. The Ombudsman's jurisdiction already needs to be significantly fixed to plug a number of holes, so that could be done at the same time. The problem of course is that pro-secrecy Ministers have no incentive to do any of this.

Subverting the OIA

A couple of months ago, the Herald reported that National MP Kanwaljit Singh Bakshi was suspected by the Immigration Service of paying off a woman at the centre of allegations he had made bogus job offers. The information - a report on the investigation from the Immigration Service - was obtained by the Herald under the Official Information Act. In response, Immigration Minister started a witch-hunt to find out why the information was released without his permission.

I was appalled by this outrageous attempt to limit OIA releases on political grounds, so I did some digging. Unfortunately, Immigration's parent body, the Department of Labour, would not release the report of Coleman's witch-hunt on a number of spurious grounds, including s6(c) (maintenance of the law) and s9(2)(e) (potential loss to the public, though given that no civil or criminal proceedings can be brought for a good-faith release under the Act, I am left wondering as to what that "loss" might be). They did, however, provide a summary [PDF], which listed the conclusions as:

The process followed was not as robust as it should have been. It was noted that there was:

• little consideration of the reputational risk in dealing with this OIA request;
• misinterpretation of the meaning of the legal advice received; and
• no management oversight of the material released.

(Emphasis added).

It goes on to recommend establishing "a process for triaging OIA requests" - i.e. identifying those which might be "politically sensitive" and ensuring that they are appropriately buried. A briefing for the Minister [PDF] expands on this:

Immigration NZ has recently put in place a further OIA triage process. This includes a panel to decide how and by which workgroup each OIA should be dealt with and to provide advice.

[...]

All OIA requests, including those received in regional offices, must be registered centrally in the Government, Executive and Ministerial Unit (GEMS).

GEMS will allocate all OIAs to the relevant 3rd tier manager.

GEMS will provide advice, assistance and monitor timelines.

(The briefing also contains an Orwellian note that the Herald reporter, Patrick Gower, has since lodged six further OIAs in the week leading up to the briefing. because obviously, the Minister needs to know that).

All of this is contrary to the spirit of the Act. As the Danks report noted way back in 1980,

The fact that the release of certain information may give rise to criticism or embarrassment of the government is not an adequate reason for withholding it from the public.

The Department's own updated OIA policy agrees with this, listing political sensitivity and the potential for Ministerial embarrassment as unacceptable reasons for withholding information. But the processes the Department has established seem to be aimed at withholding or denying information on those very grounds. And that is simply unacceptable.

Yet another attack on transparency

The Finance and Expenditure Committee reported back on the New Zealand Infrastructure Commission/Te Waihanga Bill on Friday. Its a boring bill which would establish a Crown Entity which would produce reports about infrastructure needs, so I hadn't paid attention to it. I should have, because the bill includes a - increasingly common - secrecy clause, forbidding disclosure of any information gathered using the Commission's information-gathering powers unless it is already publicly available, in statistical form, or required by an enactment. The kicker? Those information gathering powers apply only to:

1. a department named in Schedule 1 of the State Sector Act 1988, other than [the GCSB or SIS]
   
2. a departmental agency named in Schedule 1A of the State Sector Act 1988:
   
3. a statutory entity named in Schedule 1 of the Crown Entities Act 2004:
   
4. the New Zealand Defence Force.:
   
5. a local authority, as defined in section 5(1) of the Local Government Act 2002:
   
6. a council-controlled organisation, as defined in section 6 of the Local Government Act 2002 (but disregarding subsection (4) of that section):
   
7. Auckland Transport, as defined in section 4(1) of the Local Government (Auckland Council) Act 2009.
   

And disregarding some twists and turns around council-controlled port companies, every single one of those bodies is already subject to the OIA or LGOIMA. So, if they give information to the Infrastructure Commission (which as a Crown Entity is explicitly subject to the OIA), it becomes secret - but if you ask them directly, the usual rules apply.

It gets worse, because the Select Committee spent some time thinking about the OIA and LGOIMA and how it applied, tightly restricting the reasons an agency could refuse a request for information from the Commission and including a clause (currently s23(4A)) saying that information could not be withheld from the Commission if it could not be withheld under the OIA or LGOIMA. But somehow, in all of this, no-one on the Committee thought to question whether it was necessary to apply secrecy to information that was already presumptively public, and clearly no-one thought about it in the drafting process either. Our government is basicly completely blind when it comes to open government and proactive steps to protect it.

The fix for this is simple: amend s25 at the committee stage to explicitly permit disclosure under the OIA. But in the longer-term, this needs to stop happening. The government needs to think about transparency and open government when designing legislation, just as they do about human rights, the Treaty, and gender equality. Getting them to do that would make a great Open Government Partnership commitment.

(And meanwhile, there's another one: the new Venture Capital Fund Bill allows the (Crown Entity, subject to the OIA) Guardians of New Zealand Superannuation to create "VCF investment vehicles" under their control which are explicitly exempted from the OIA. This runs counter to the usual control principle, under which government-controlled entities such as subsidiaries of CRIs, SOEs and Crown Entities (and local government ones like CCOs) are subject to the OIA and/or Ombudsmens Act. No policy justification is provided for this secrecy).