I wonder whether the assertion of undersupply is a reflection on "affordable" (i.e. lower price) rentals available, rather than rentals overall. Not sure that it's possible (whether data or person time), but I wonder what the 2.5% would look like in different e.g. market quartiles. If all the over-supply is in upper quartile apartments, that's unlikely to quickly trickle down into improving supply in lower quartile family homes?
I agree it would be good to analyse by price, but I cannot create that back data. I could not confidently predict your guess. I do run charts by district for Wellington, which is a bit of a proxy for price, inventory in Porirua (50%) and Lower Hutt (50% up) is also high, not quite for Upper Hutt (only higher 2 of 8 years). You can see that data grouped in my reference charts above).
As an investor in multiple properties over 20 years, all in premium markets, we always meet the market for price so if all the oversupply was in unaffordable areas, those rentals would become affordable very quickly. I have never left a price longer than 2 weeks before dropping it to meet the market. Trickle down is actually how the market works, investors can never dictate prices.
I do assert that regularly, so I wasn't really happy to be suggesting otherwise ;)
However, presumably there's some time involved in the ebb and flow across the market? e.g. if there was 3.0% vacancy rate in the upper quartile, 2.5% in second and third quartiles and 2.0% in the lowest quartile, it might take a wee while for the market to adjust to 2.5% across the board?
Measuring by rental quartile would be tricky, but by number of bedrooms is a bit easier. My denominator is different to yours (not sure how/why the quarterly data active bond numbers are so different to tla-level monthly data), but it makes the point anyway: if I take active bonds for SA2 areas in Wellington City in October 2023 (from the detailed quarterly bond data), and today's Trademe rental listings, it gives vacancy rates of 4.3% (251/5796) for 1 bedroom properties, 4.9% (278/5649) for 2 bedrooms, 4.4% (254/5808) for 3 bedrooms, 11.3% (139/1227) for 4-bedrooms, and 14% (71/507) for 5+bedrooms. So even though the total vacancy rate is 5.2% (993/18987), it looks quite different from that in different parts of the market. Those landlords renting out 4 and 5+bedroom properties could presumably rent to people who are looking for 3 bedrooms if they dropped their price, but it doesn't look like they're doing that in a hurry? Can only assume not all landlords are as sensible as you :D
Possibly they can't stomach the amount of price drop necessary to make someone who only needs 3 bedrooms take a 4-bedroom instead; and those who can afford upper quartile 3-bedrooms will mostly choose a better quality 3-bedroom over a price-reduced 4-bedroom?
Yea but your thinking encourages me and others to think more too. Keep it up.
Many of the multiple bedroom homes are group (student) rentals so are often effectively priced by the room (the CBD rents are centred on $200-300), raising the total price but potentially lowering the vacancy rate. First tenants are expected to find flatmates themselves on a different page, (a market we quickly exited 20 years ago due to the costs of social management when the flat went wrong). With first tenants, landlords will not lower their prices for good economic reasons - I wish many would raise their quality but that's a different story.
Australian cities are short on rentals according to their Realestate.com.au who also use the 3% figure but report numbers very close to 1%. I've been looking for a rental in Melbourne and the market seems to be running normal (2 week turnaround), but Realestate are saying it's hot, as they would. I will look for better data for that since it would be useful to compare.
On your vacancy rates, that requires a lot of unpicking. I use the same raw data but aggregated for the whole region for my vacancy rate comparison between regions. I have one inventory chart covering just the Wellington/CBD area, and it shows very high inventory but I've not extracted the associated active bonds. Your numbers suggest the lower inventories to get as low as 3% region-wide must be found in outer suburbs - or turnover rates are very different in outer suburbs. A whole lot of more research!
Hi John, thanks for sharing this work you do. Question: how have you factored in population increases/decreases in these areas over the years into your over/under supply?
Thank you for your question Conor, I hope I answer it fully for you and others, its a good question.
Population change is effectively automatically included, eg Geometric mean rents/household income. Geometric Mean rents is from Tenancy Services and is a mean of all house rents, so an increase in population increases the number of rents included in the mean. Household Income is published annually in June and actually has the number of households, but again that is not going to change the median household income that I use , which is the sum of all incomes divided by the number of households - again a population ratio
The other main data source I use is the number of rental properties advertised each week and of course that is also population dependent, however I also divide this by household income to get a vacancy rate (target 2.5% balance) so population ends up in both the nominator and denominator cancelling out.
When I quote the Geometric Mean Rents, that is a median price for rental properties released by the RBNZ later but initially by MBIE. They state it is an average taking account of the number of bedrooms ie 2,3,4... in each region so it can be assumed accurate as an average rent although they also publish the median which I use, removing very high or very low rents to some extent in each region to get the Geometric Mean.
I do not yet use advertised vacancies with population, but that shows in the graphs, ie as population grows I would expect vacancies to grow in tangent, ie each year the lines should be higher than the previous year, but they don't due to other influences - see those charts here https://johnbutt.substack.com/p/inventory-charts
Fascinating as always John,
I wonder whether the assertion of undersupply is a reflection on "affordable" (i.e. lower price) rentals available, rather than rentals overall. Not sure that it's possible (whether data or person time), but I wonder what the 2.5% would look like in different e.g. market quartiles. If all the over-supply is in upper quartile apartments, that's unlikely to quickly trickle down into improving supply in lower quartile family homes?
I agree it would be good to analyse by price, but I cannot create that back data. I could not confidently predict your guess. I do run charts by district for Wellington, which is a bit of a proxy for price, inventory in Porirua (50%) and Lower Hutt (50% up) is also high, not quite for Upper Hutt (only higher 2 of 8 years). You can see that data grouped in my reference charts above).
As an investor in multiple properties over 20 years, all in premium markets, we always meet the market for price so if all the oversupply was in unaffordable areas, those rentals would become affordable very quickly. I have never left a price longer than 2 weeks before dropping it to meet the market. Trickle down is actually how the market works, investors can never dictate prices.
I do assert that regularly, so I wasn't really happy to be suggesting otherwise ;)
However, presumably there's some time involved in the ebb and flow across the market? e.g. if there was 3.0% vacancy rate in the upper quartile, 2.5% in second and third quartiles and 2.0% in the lowest quartile, it might take a wee while for the market to adjust to 2.5% across the board?
Measuring by rental quartile would be tricky, but by number of bedrooms is a bit easier. My denominator is different to yours (not sure how/why the quarterly data active bond numbers are so different to tla-level monthly data), but it makes the point anyway: if I take active bonds for SA2 areas in Wellington City in October 2023 (from the detailed quarterly bond data), and today's Trademe rental listings, it gives vacancy rates of 4.3% (251/5796) for 1 bedroom properties, 4.9% (278/5649) for 2 bedrooms, 4.4% (254/5808) for 3 bedrooms, 11.3% (139/1227) for 4-bedrooms, and 14% (71/507) for 5+bedrooms. So even though the total vacancy rate is 5.2% (993/18987), it looks quite different from that in different parts of the market. Those landlords renting out 4 and 5+bedroom properties could presumably rent to people who are looking for 3 bedrooms if they dropped their price, but it doesn't look like they're doing that in a hurry? Can only assume not all landlords are as sensible as you :D
Possibly they can't stomach the amount of price drop necessary to make someone who only needs 3 bedrooms take a 4-bedroom instead; and those who can afford upper quartile 3-bedrooms will mostly choose a better quality 3-bedroom over a price-reduced 4-bedroom?
I'm mostly thinking as I type here, mind you.
Yea but your thinking encourages me and others to think more too. Keep it up.
Many of the multiple bedroom homes are group (student) rentals so are often effectively priced by the room (the CBD rents are centred on $200-300), raising the total price but potentially lowering the vacancy rate. First tenants are expected to find flatmates themselves on a different page, (a market we quickly exited 20 years ago due to the costs of social management when the flat went wrong). With first tenants, landlords will not lower their prices for good economic reasons - I wish many would raise their quality but that's a different story.
Australian cities are short on rentals according to their Realestate.com.au who also use the 3% figure but report numbers very close to 1%. I've been looking for a rental in Melbourne and the market seems to be running normal (2 week turnaround), but Realestate are saying it's hot, as they would. I will look for better data for that since it would be useful to compare.
On your vacancy rates, that requires a lot of unpicking. I use the same raw data but aggregated for the whole region for my vacancy rate comparison between regions. I have one inventory chart covering just the Wellington/CBD area, and it shows very high inventory but I've not extracted the associated active bonds. Your numbers suggest the lower inventories to get as low as 3% region-wide must be found in outer suburbs - or turnover rates are very different in outer suburbs. A whole lot of more research!
My last two rentals are on the market for sale as we exit PI for retirement purposes - https://www.trademe.co.nz/a/property/residential/sale/wellington/lower-hutt/eastbourne/listing/4665637331 and its other half.
Hi John, thanks for sharing this work you do. Question: how have you factored in population increases/decreases in these areas over the years into your over/under supply?
Thank you for your question Conor, I hope I answer it fully for you and others, its a good question.
Population change is effectively automatically included, eg Geometric mean rents/household income. Geometric Mean rents is from Tenancy Services and is a mean of all house rents, so an increase in population increases the number of rents included in the mean. Household Income is published annually in June and actually has the number of households, but again that is not going to change the median household income that I use , which is the sum of all incomes divided by the number of households - again a population ratio
The other main data source I use is the number of rental properties advertised each week and of course that is also population dependent, however I also divide this by household income to get a vacancy rate (target 2.5% balance) so population ends up in both the nominator and denominator cancelling out.
When I quote the Geometric Mean Rents, that is a median price for rental properties released by the RBNZ later but initially by MBIE. They state it is an average taking account of the number of bedrooms ie 2,3,4... in each region so it can be assumed accurate as an average rent although they also publish the median which I use, removing very high or very low rents to some extent in each region to get the Geometric Mean.
I do not yet use advertised vacancies with population, but that shows in the graphs, ie as population grows I would expect vacancies to grow in tangent, ie each year the lines should be higher than the previous year, but they don't due to other influences - see those charts here https://johnbutt.substack.com/p/inventory-charts