Here are the latest data for Auckland dwelling consents. I've included the 'actual' data and the 'trend' data': the 'trend' version is Stats' best effort to abstract from the (quite considerable) month to month volatility and to show us the underlying picture. I've gone back to 1995, partly because that's where the 'trend' series starts in Stats' database and partly to put the current rate of building into context.
It's good news as far as it goes. That dip has gone away, and it's onwards and upwards in recent months. It's still not clear why we had that earlier dip: some people I've spoken to said that developers were waiting to see the shape of the Auckland Unitary Plan, and maybe that's true. But it's somewhat at odds with the recent rises, which predate the publication of the Plan (it went public on July 22 and was only signed off by the Council on August 19). Perhaps there'll be another hiatus as the Plan is appealed, or maybe developers aren't fixated on the Plan at all: we'll have to wait and see.
Another possibility is that the trend-detecting algorithm at Stats had a temporary hissy fit and is now back on track, though the downside of that thought is that if it detected dips when there weren't any, maybe it's finding surges when there aren't any, either. And yet another possibility is that the month to month numbers (especially for apartment blocks) are just too volatile to find a reliable signal in all the noise, even if your trend-spotting software is up to scratch.
But in any event, let's bank it - as far as it goes, which isn't far enough. You can see for yourself that this recent rate of consenting is getting closer but still isn't back up to the levels of the early 2000s, and on a per capita basis it's still well adrift: urban Auckland's population was around 1.2 million in the early 2000s, and it's nearer 1.5 million today.
And it's not just an Auckland problem, either: nationally we seem to have had increasing difficulty in getting homebuilding activity up to the level it needs to reach. This chart (from Stats' info release today) shows that we are currently consenting some 30,000 dwellings a year (and that's inflated quite a bit by the Canterbury rebuild), but we were regularly clocking 30,000 and more in the early 1970s when the population was only around 3 million compared to today's 4.7 million. There are more sophisticated ways of measuring it, but I doubt if they'd shake the basic conclusion, which is that we need to get the supply side of the market operating a good deal more responsively.
I hadn't been following the Canterbury rebuild numbers closely: I had a vague impression that peak rebuild was somewhere around now. And indeed it is, on these figures from the June Building Activity Survey: the total value of building work in the Canterbury region is still going up a bit, but looks like it is plateauing.
Within the overall total, the housing rebuild is actually past its peak, which was back in late 2014 and early 2015, while the non-housing rebuild is still growing.
Today's dwelling consents figures for Canterbury show the same picture: consents also peaked back in the second half of 2014.
There's still a lot of housebuilding going on in Canterbury, and self-evidently the job can't be finished, but hopefully the recent modest drop in activity is a signal that the bulk of the demand has been met and there's less left in the pipeline still to do.
The other thought that emerges from these data is that as far as the impact on overall GDP growth is concerned, the boost from the Canterbury rebuild has largely run its course. Ideally the construction sector would swing more or less smoothly from meeting demand in Canterbury to meeting demand in Auckland, though as noted above our national ability to meet demand has become progressively creakier. We could even see further GDP growth if the Auckland build started to become larger than Canterbury's. But overall it's beginning to look as if we're going to find something else to do with our resources if we want to keep 3% growth going.