jesuslover07
New member
(Note: I made a post similar to this yesterday, which I deleted due to a gratuitous
error. Thanks to /@birch for pointing out the mistake!)
I did some rough calculations, and as a result I'm becoming more convinced that the
forecasts of 10-15% drops in housing prices are pretty reasonable.
I already leaned towards taking folk like Chris Joye at their word that a 15-25% drop
was reasonable, based on the bloke talking plenty of sense about how he arrived at that
number, and his being a verified non-permabear - previously arguing correctly that
prices would rise during the pandemic. This gives him a lot of credibility in my eyes.
And most banks are making similar projections. My own numbers here are for a somewhat
smaller drop - especially since Joye's forecast is assuming only a 100bp rate increase
and we've already exceeded that. But it's not crazy different. Maybe I'm a tad more
optimistic with the assumptions.
Anyway. The core fact is that prices are limited by loan serviceability.
Yearly principal+interest mortgage repayments on an
-year loan of size
with interest rate
, ignoring
intra-year compounding are:
Assuming annual repayments for which someone meets serviceability requirements are
proportional to income
, and assuming loan size is proportional to the price
,
substituting
and
into the above equation and solving for
gives us a dumb
model of how prices might vary with income and interest rates:
The rate
in this equation is the rate used for serviceability tests - which would be
3% above the mortgage rate as per APRA requirements.
Assuming a 30-year loan, that mortgage rates are 2.4% above the cash rate and that peak
prices occurred when the cash rate was 0.1%, we can estimate price drops from peak in
various scenarios.
Assuming the cash rate increases to 2.85% (the median
forecast
of surveyed economists), and that nothing else changes (no income growth), this
oversimplified model says prices would fall by 24% from peak.
Sounds like a big drop!
However, some other things likely will change. For one, if that takes two years to play out,
then nominal income growth might be something like 7% (the resulting growth in house
prices won't be real, but we're talking nominal here). Also, it seems likely that APRA
will decrease the serviceability buffer from 3.0% to 2.5% once rates get a bit higher.
They might even decrease it more.
Take those assumptions into account and you get that prices will fall 15%.
And, 2.85% is the peak cash rate forecast by economists. By end of 2023 the median
forecast is 2.6%, which if fed to the above, gives a price fall of only 13%.
You can get more optimistic if you like: Rates won't necessarily stay where they are by
end of 2023. The CBA thinks rates will decline to 2.1%. If that happens, then when all
is said and done the drop in house prices would be only 9%.
So there's a range of scenarios there that I think reinforce the expectation of a drop
of 10-15%. You can be a bit more pessimistic and get it to a 24% drop if you like, but
you have to assume zero nominal wage growth and APRA holding their buffer constant. And
you can get a bit more optimistic and assume the CBA's forecast is on the money and get
it to a 9% drop. All in all 10-15% ends up looking like like a pretty reasonable
central expectation.
One thing ignored here is that deposits are not a fixed percentage of sale price: when
prices are lower, deposits are more likely to be a larger fraction of the sale price.
This helps serviceability and so will result in somewhat smaller price drops than the above model
says.
And of course it's a dumb and simple model that ignores everything else that would
affect demand for and supply of housing. It's just talking about the ability to get
credit and even then is super oversimplified.
Here's a sample of model outputs including the ones mentioned above, and some others:
error. Thanks to /@birch for pointing out the mistake!)
I did some rough calculations, and as a result I'm becoming more convinced that the
forecasts of 10-15% drops in housing prices are pretty reasonable.
I already leaned towards taking folk like Chris Joye at their word that a 15-25% drop
was reasonable, based on the bloke talking plenty of sense about how he arrived at that
number, and his being a verified non-permabear - previously arguing correctly that
prices would rise during the pandemic. This gives him a lot of credibility in my eyes.
And most banks are making similar projections. My own numbers here are for a somewhat
smaller drop - especially since Joye's forecast is assuming only a 100bp rate increase
and we've already exceeded that. But it's not crazy different. Maybe I'm a tad more
optimistic with the assumptions.
Anyway. The core fact is that prices are limited by loan serviceability.
Yearly principal+interest mortgage repayments on an
Code:
N
Code:
L
Code:
r
intra-year compounding are:
Code:
annual_repayments = r L / (1 - (1 + r)^-N)
Assuming annual repayments for which someone meets serviceability requirements are
proportional to income
Code:
I
Code:
P
substituting
Code:
I
Code:
P
Code:
P
model of how prices might vary with income and interest rates:
Code:
P ∝ I / r × (1 - (1 + r)^-N)
The rate
Code:
r
3% above the mortgage rate as per APRA requirements.
Assuming a 30-year loan, that mortgage rates are 2.4% above the cash rate and that peak
prices occurred when the cash rate was 0.1%, we can estimate price drops from peak in
various scenarios.
Assuming the cash rate increases to 2.85% (the median
forecast
of surveyed economists), and that nothing else changes (no income growth), this
oversimplified model says prices would fall by 24% from peak.
Sounds like a big drop!
However, some other things likely will change. For one, if that takes two years to play out,
then nominal income growth might be something like 7% (the resulting growth in house
prices won't be real, but we're talking nominal here). Also, it seems likely that APRA
will decrease the serviceability buffer from 3.0% to 2.5% once rates get a bit higher.
They might even decrease it more.
Take those assumptions into account and you get that prices will fall 15%.
And, 2.85% is the peak cash rate forecast by economists. By end of 2023 the median
forecast is 2.6%, which if fed to the above, gives a price fall of only 13%.
You can get more optimistic if you like: Rates won't necessarily stay where they are by
end of 2023. The CBA thinks rates will decline to 2.1%. If that happens, then when all
is said and done the drop in house prices would be only 9%.
So there's a range of scenarios there that I think reinforce the expectation of a drop
of 10-15%. You can be a bit more pessimistic and get it to a 24% drop if you like, but
you have to assume zero nominal wage growth and APRA holding their buffer constant. And
you can get a bit more optimistic and assume the CBA's forecast is on the money and get
it to a 9% drop. All in all 10-15% ends up looking like like a pretty reasonable
central expectation.
One thing ignored here is that deposits are not a fixed percentage of sale price: when
prices are lower, deposits are more likely to be a larger fraction of the sale price.
This helps serviceability and so will result in somewhat smaller price drops than the above model
says.
And of course it's a dumb and simple model that ignores everything else that would
affect demand for and supply of housing. It's just talking about the ability to get
credit and even then is super oversimplified.
Here's a sample of model outputs including the ones mentioned above, and some others:
Code:
Cash rate to 12.0%, all else equal: -60.8%
Cash rate to 3.50%, all else equal: -28.7%
Cash rate to 2.85%, all else equal: -24.3%
Cash rate to 2.60%, all else equal: -22.5%
Cash rate to 2.10%, all else equal: -18.7%
Cash rate to 12.0%, income +7%: -58.0%
Cash rate to 3.50%, income +7%: -23.7%
Cash rate to 2.85%, income +7%: -19.0%
Cash rate to 2.60%, income +7%: -17.1%
Cash rate to 2.10%, income +7%: -13.0%
Cash rate to 12.0%, income +7%, APRA buffer to 2.5%: -56.8%
Cash rate to 3.50%, income +7%, APRA buffer to 2.5%: -20.2%
Cash rate to 2.85%, income +7%, APRA buffer to 2.5%: -15.1%
Cash rate to 2.60%, income +7%, APRA buffer to 2.5%: -13.0%
Cash rate to 2.10%, income +7%, APRA buffer to 2.5%: -8.6%