Te Kawa Mataaho / Public Service Commission has released its half-year OIA statistics today. Last year they received some harsh media coverage calling them "close to useless" due to various rorts, and in response this set includes information on extensions, transfers, and response times. Public Service Commissioner Peter Hughes is trumpeting that the average response time across all agencies other than police and NZDF was 12.5 working days, and saying that this shows that everything is fine and "the use of extensions, transfers and refusals [is] not a factor". It doesn't. Instead, the focus on the average performance is being used to obscure some rather significant failures.
Firstly, the good news: three quarters of the 32 core agencies met the expected standard of handling 95% of requests on-time (and two of the ones which didn't had a single late request). Some of these agencies are doing very well: seven of them had median response times of less than ten working days. And three agencies - Corrections, MPI, and Customs - had a median response time of only two days. Which is genuinely good, but also suggests that these agencies have a large number of routine, standardised requests which are processed quickly, and that the performance for "real" OIAs sent to their head office may be different. And that's actually supported by MPI's data, where the two day median is combined with a 15.4 day average, suggesting a bimodal distribution where roughly 2/3rds of their requests are routine and quickly processed, while the rest take the full 20 working days or longer (Ministry of Health shows a similar pattern, combining a six working day median / eight working day average with an on-time rate of only 90%, which works out pretty nicely if you assume all late requests take only 21 days).
Secondly, two of the top performing agencies - Customs (average 5.7 days) and Ministry of Health (average 8 days) are two of the busiest in the public sector, and largely determine the low overall average response time. Meanwhile, ten agencies - including major media request targets such as MBIE, DPMC, MSD and Ministry for the Environment - have average response times of more than 20 working days, while 12 have a median response time of 20 or more days (Oranga Tamariki's is 21, FFS). Which is a clear sign that they're treating the OIA's 20 working day time limit as a target rather than a limit, and pissing on the requirement to respond as soon as reasonably practicable.
As noted in that Stuff article above, long response times don't necessarily mean requests are late, as agencies can extend. And contrary to Hughes, there is clear evidence that some agencies are rorting the system to juke their stats. Sorting the table by extension percentage shows us that a typical agency extends less than 20% of requests. Meanwhile, six agencies are extending more than a quarter of requests, including MfE, which extends more than half, MFAT (44%), and DPMC (34%). These agencies are rorters, and now we can all see it (the spies are also in this category, and have recently been spanked for it by the Ombudsman).
Astonishingly, two of the worst extension rorters - the spies and DPMC - are also the worst performers for (legal) timeliness. MBIE is also a poor performer, with only 91% on time. As for non-public service departments, the police claimed a 98% on-time rate by rorting their stats, while NZDF was late almost 50% of the time and refused to provide response time statistics. Which sounds like they have contempt for the law and are in severe need of a practice investigation by the Ombudsman.
All of which explains the derision Hughes' sunny proclamation was met with from journalists and requesters. There's a story in his new, improved statistics, but he's simply not interested in telling it. Which also means that he's not interested in improving OIA performance, and is content to let the rorters juke the stats without penalty. Whether this is fulfilling his statutory duty to provide leadership and promote transparency and accountability in the public service is left as an exercise for the reader.
