Okay, what you are planning sounds good, Statistics NZ - but we still need more.
Sorry to sound ungrateful but that's my immediate reaction to news that our crown-controlled statistical information provider is moving to provide more timely pricing data.
To simplify enormously what is proposed, at the moment Stats NZ produces monthly pricing information (on food and rent) that comprises about 30% of the make up of the Consumers Price Index - our accepted measure of inflation. From November 14 Stats NZ will be expanding this to include further items, the upshot being that we'll now get a monthly snapshot of pricing information that makes up about 44% of the CPI.
Well, that's better, of course. But how about as close to 100% of that CPI information as possible, every month? That would align us with other countries - Australia, for example which produces a monthly 'CPI Indicator' alongside the fully quarterly CPI information.
We should do this too. Absolutely.
And I will add a further one in here, which is very much related - what about GDP figures?
Like the CPI data, our GDP information is produced quarterly. But the GDP info is nothing like as up to date by the time we see it.
The CPI data typically comes out say two-and-a-half to three weeks after the end of the quarter.
But the GDP data doesn't appear till damn near the end of the NEXT quarter.
As an example, we got our eyes on the September quarter CPI figures in this past week - but we won't see September quarter GDP figures till nearly midway through December, at which point most eyes will be on the Christmas break.
Does our lack of timely data of this nature make us unusual? Yes, it does. As the IMF pointed out in its annual review of New Zealand this year:
"The lack of a monthly CPI series makes New Zealand an outlier among advanced economies and is holding back a timelier formulation and assessment of monetary policy."
The Reserve Bank has conceded it didn't react fast enough to the post-pandemic explosion in inflation. Would more timely information have helped?
It's worth just having a quick look back to mid-2021. The March quarter CPI had shown us having an annual inflation rate of 1.5%. Remember the RBNZ's inflation target is a range of 1% to 3%. So, 1.5% is just fine and dandy.
The March quarter 2021 figures came out in April and then in late May the RBNZ produced its Monetary Policy Statement in which it forecast annual inflation as of the June quarter 2021 - that we were almost two thirds of the way through by that stage - of 2.6%, so yes, up a fair bit. But it then forecast inflation to FALL in subsequent months.
Anyway, as the time for release of the June quarter CPI figures got closer, the country's bank economists were having their picks and they reckoned 2.8% to 3% could be the figure. July 16 told the actual story. It was 3.3%. All hell broke loose. That CPI release was actually two days AFTER the RBNZ had reviewed the OCR and left it UNCHANGED on the emergency setting of 0.25%.
It will show as a historical anomaly that the RBNZ did NOT hike the OCR at its next review in August either. But that was solely because confusion reigned as Auckland experienced its outbreak of Covid Delta the day before the RBNZ review! The hikes duly commenced in October 2021, stopping only (at least for now) in May of this year with the OCR at 5.5%.
Now, I'm not trying to say that monthly CPI inflation would have enabled the RBNZ to magically stop inflation in its tracks - but when you are dealing with what very soon became a runaway train (the inflation, I mean), every moment counts. The longer a country has high inflation the more the risk it becomes seriously ingrained as the population becomes 'accustomed' to it and incorporated inflation into their future (pricing) decisions.
Our annual rate of inflation, we found out in the past week, dropped to 5.6% as of the September quarter, down from 6.0%. That's the first time since December 2021 - getting on for two years - that we've had an annual inflation figure without either a '6' or a '7' at the front of it. Time will tell whether we can get inflation back into its 1% to 3% box.
So, anyway, time for some more timely data. I reckon we need a monthly CPI Indicator and also some form of 'early' GDP indicator that comes out as soon as possible after the quarter in question finishes. But how do we do this?
National Party deputy leader and finance minister presumptive Nicola Willis has frequently challenged RBNZ Governor Adrian Orr on our rates of inflation being higher than other countries and slower to fall. Well, she is about to be in a position to do something about it.
The obvious question is, what resources would Stats NZ need to be able to produce this more timely information?
As per its 2022 annual report, Stats NZ has about 1500 staff.
In the 2023-24 Budget, Vote Statistics was allocated just under $266 million, which included nearly $45 million to finish off the 2023 Census. There was also around $30 million for capex and services to other agencies.
And then the remainder, of a little over $191 million was allocated thus:
• nearly $69 million for the delivery of data and statistical information services relating to business and the economy
• over $84 million for delivery of data and statistical information services relating to the population, household economics, social conditions (including child poverty), the labour market, and the environment
• just over $38 million for the coordination of statistical and data services for government, through System Leadership of the Official Statistics System (OSS) and Stewardship of the Government Data System; statistical and data management advice; provision of access to official statistics; oversight of the IDI; and the provision of ministerial services
As a matter of interest, in the last National Government Budget for 2017-18, this was the comparable breakdown of a total of just over $121 million:
• A total of nearly $52 million for the delivery of data and statistical information services relating to business and the economy.
• A total of just under $42 million for the delivery of data and statistical information services relating to population, environment, household economics, social conditions, and the labour market.
• A total of just under $28 million for system leadership for data and analytics across the public sector, oversight of the Integrated Data Infrastructure (IDI), statistical and data management advice and the operation of access channels.
Based on those figures there does appear to have been quite a switch during the Labour Government in terms of resources put toward what could be loosely described as 'social' data and away from the business and economic data.
I'm a data nerd. For me, the more data the merrier about anything. But what is the new National-led Government likely to think?
It's quite possible there would be a push to adjust priorities.
Watch this space.
But whatever, I would just like to see Stats NZ given the resources to produce timely information of the type I've outlined in this article.
Our key economic planners would be grateful and, ultimately, so should we be. This type of information is not a 'nice to have' - it is a 'must have'.
*This article was first published in our email for paying subscribers early on Friday morning. See here for more details and how to subscribe.
14 Comments
Back when I was an analyst programmer one of my first jobs was writing a suite of reporting programs that would be run monthly, quarterly and annually.
These programs would drill into the transactions created in the previous periods and aggregate them up. This seemed to me to be an inefficient and expensive way to do do this.
Surely a better way would be to place the summary of the transaction onto a queue with the queue being processed asynchronously in near-real time so that the aggregate values were maintained continuously?
Some years (late 80s) later I got a chance to pilot this for a very large financial institution processing 100s of transactions per second. It worked a treat. Especially as the occasional transactional errors that delayed the monthly, quarterly and annually reports were recognised and corrected immediately. (Huge operational gains in that. Massive in fact. Way, way more saved over just 6 months than my system cost to develope.)
All sorts of issues were raised about why my approach couldn't work were tabled. (What if there was an error in a transaction being one of them.) None of these blockers turned out to be real issues. (The errors in the data issue was dealt with by the near-real reports showing "provisional" vs. "actual" when the results had been validated.) Legal and compliance issues were raised continuously, mainly by people who had no idea what they were talking about. None were actually valid, or couldn't be solved by restricting access to the data on a "need to know" basis.
This "near-real time" paradigm is pretty much standard practice nowadays, albeit there are many competing technical approaches that ICT folk love to argue over.
Methinks, Stats NZ processes need some attention. ;-)
I like it (and are also building similar where I currently work) but wouldn't quite work in Stats, cos you would be mixing data sets that come in at different times, run by different teams (I also used to work at Stats in IT). So it's May and the team feeding in one set of data that makes up 10% of the CPI feeds in their data for April. But you only get the April data from another team that represents another 10% in 5 days time. So for the intervening period, you are mixing older data with newer, if that's published, you are getting poor data quality. The Statisticians would be pulling their hair out, especially so because other organisations use the data to decide actions (RBNZ just being one). Include in this potentially limited checks (so possibility of high errors) and you basically would be struggling to ever be happy about the figures being reported. That's why they have a release process where data from a dozen different teams is collated and reviewed by the Senior Stats nerds, then released. But no reason this can't be done more regularly.
It would work fine at Stats NZ.
Every transaction has two notable dates. The date at which the transaction occurred. And the effective date. The effective date is when the transaction occurred and is separate to date at which the the transaction becomes effective. The collection and processing of the individual items that teams are working on are no different. The "price of a can of bean for November" is a transaction in this context with the transaction date being when the price was collected but November is the effective date.
In the 'publishing' context you refer to, there is also a publishing date at which the transaction becomes published (and this could be the effective date). Although 'published' in this context is in effect external to the source transaction and would be subject an authourization scheme so that some users can see stuff while others can't. (The "need to know" context I mentioned.)
Their "release process" can be tracked back to paper systems, and then to disparate systems collecting and cleaning data before publication. A very old, traditional (and insecure and error fraught) process. If the data must all be published as a single release, then the underlying data (the transactions) can be stored in a single data source. Who sees what, when, and how of this data is nothing but a "view into the data".
Thus when Stats NZ "publish" it's simply a flick of a switch that says, "This group can see the final result now".
We're in California at the moment and have been here for 20 days.
We are very slow to adapt on many fronts such as,
Some parking lots closed between 11pm and 4am, must have a permit from the police, probably stops some unlawful activity.
No permit to park in a mobility park, $250 fine.
Hand wash stations in the same location as portable toilets at events.
As a country we have become too soft on people who don't follow societies rules.
If other countries can produce inflation and other financial stats quicker, why can't we.
Systemic slackness.
The number one priority for the Stats department is figuring out how many Maori people there are. That about sums up the Labour Govt.
https://www.1news.co.nz/2023/07/12/stats-nz-to-release-first-census-dat…
Layer upon layer upon layer (of middle management that is). Reality eventually caught up with Sara Lee. The makers of that famous 80's commercial. Now Stats and many other government departments are going to get bitten by reality very soon. Anyone with any connection to international shipping would have seen inflation coming form 5000 miles away in 2020. The bureaucrats had no idea because they live in a Wellington beltway bubble. Looking backwards, revising data that is already 6 months old.
I've been calling this out for 2-3 years or more. You can check out my previous posts on this site & others as well. Most of my news comes from offshore where stat/state data is available way quicker than it is here. It is symptomatic of our central governmental decline across the board. We keep having to look backwards for the reality when we are already 2-3 months into the new world. Too late state people! We have the technology. it's out there. Find it & use it.
We welcome your comments below. If you are not already registered, please register to comment.
Remember we welcome robust, respectful and insightful debate. We don't welcome abusive or defamatory comments and will de-register those repeatedly making such comments. Our current comment policy is here.