[tmnvix]

I can confirm this in the context of NZ.

During the past 25 years the number of residences relative to the number of households has increased substantially from roughly 3.5% more residences to something like 6.5% more. All while the number of people per household has declined slightly and prices have increased astronomically.

Very frustrating this past decade to have the entire debate framed around building more houses as the only solution to deal with this 'undersupply'.

In my opinion a speculative boom has created excess demand and increased the number of underutilised properties. Thankfully rising interest rates appear to be dampening that excess demand (to put it mildly).

[thelamest]

What you wrote doesn’t appear to be true.

Divide columns B7 on sheets 2 and 3 in the NZ data (https://www.stats.govt.nz/information-releases/dwelling-and-...). The all-time high of 1.06 happened in 2012 and the ratio has since fallen back down to around 1.043, just a bit above the 1.04 value from the 1990s.

My prior is that there is an undersupply, signalled by prices, and I’m afraid this method tends to underestimate it. First, household formation (as in, moving out and starting a family etc.) depends in part on housing availability. Second, internal migrations due to urbanization cancel out in national level data – for every move to a city, there may now be an empty rural unit, and a crowded urban unit.

PS Remember too, that there’s a natural churn to housing – people sell/renovate/etc. on average once every X months – therefore we need at least 1/X of excess housing to smooth it out. This is an absolute lower bound that assumes people don’t differentiate between any two houses, but are willing to randomly swap e.g. a 120 sqm in the central business district for a 30 sqm in a rural area. This assumption is obviously wrong.

[tmnvix]

You are right. Those figures are off but they are based on the same dataset you are referring to (or at least an earlier version).

I notice that they have revised some of the more recent household estimates. I had 1991 at 1,307,000 private dwellings and 1,252,600 households. A difference of 4.16%. And 2017 (or rather Q4 2016) at 1,855,500 private dwellings and 1,734,800 households. A difference of 6.5%. This - at least a few years ago when I first looked this up - was the most recent period where both figures were available.

They have indicated the revision to the household data and I am almost certain that they have revised the dwelling data as well.

The long and the short of it is, we still have a higher ratio of dwellings to households than we did before this price boom started to take off.

If I am reading you correctly, and you are indeed saying that rising prices are enough to assume an undersupply, I would have to both agree and disagree. I think it is only enough to indicate an undersupply relative to total demand, but that demand is both for homes and investments. As interest rates fall, monthly mortgage payments stay roughly the same while the size of mortgages rises as people are able to bid more for a house. With rising prices comes speculation and extra demand. We are about to find out what happens when that type of demand is removed or at least severely curtailed.

Your point about churn is definitely true. I mentioned in another comment in this thread that 'on the market' is one of many reasons that a house can be empty. I take your point but also believe that turnover is more frequent when speculation is at play.

If you are interested in measuring supply/inventory/utilisation of residential property, I'd recommend that you look up the 'Speculative vacancies report' published by Prosper Australia, a Georgian group based in Melbourne. They have an interesting methodology whereby they gather water usage data and use that to determine if a property is underutilised/vacant.