Analysis of potential house price growth long term

Hello

I’ve been sitting on this for a little while but finally got around to putting it together. Apologies as it may get a bit long.

What i’d like to do below is provide a few ideas on the outlook for house prices in mid to long term based on house price drivers, current valuations and policy settings, and what financial markets are currently pricing in for Australian interest rates.

Before I go further, a few caveats:

• All the data is based on national averages. I realise there will be large variations between cities, suburbs etc, and also that individual investors have the capacity to achieve results that vary significantly from local let alone national market averages.
• The data in the model is fairly reliable (bar a few annoying changes in definitions and calculation methodologies by the RBA and ABS over the years). I’m sure many will have got different rates from banks to what’s assumed in the model. That’s not really the point. Its about broad national averages and consistency over time. Please don’t get caught up on the little details.
• A model is just a model. And this one is not predictive. It goes some way to explaining why house prices have done what they have done over extended periods, and at the moment what might be possible in the future, not it is not a forecasting model as such.
• But I’m not trying to make any specific case, just walk through the data to stimulate discussion among a community that seems to include plenty of experienced property people. Some assumptions here do catch my eye (like people assuming 7-10% p.a. growth long into the future), so some of this may be useful for some people. But please take in the spirit in which it is intended. I’m also sure many investors couldn’t care less about national averages etc. I’ve got no problem with that and wish you the best.
• Full disclosure: I’m not anywhere near as much into property as I would like, and would be financially much better off if the market rolled over tomorrow in a big way to provide some cheaper entry points.
• Despite the above, I’ve made a pretty good career so far out of doing financial analysis.

The model

The valuation model I’m using is the one below. I’m sure a number of you have seen some version of it elsewhere. The picture below is a bit outdated now, but all I could find publicly available. But I’m quite familiar with its inputs, and have created my own fully updated version. The model basically calculates an estimated national average house price based on mortgage rates and household incomes. As you can see, from 1980 to 2014 it did quite a good job of explaining price growth, and I can say that it has proved equally as accurate at predicting the trend in house prices since then.

Over time the model has been pretty good.

• Over the 80’s and 90’s, based on household income growth of 5.9% p.a. and a 0.9% p.a. boost in purchasing power from mortgage rates falling from 10.05% to 7.27% (albeit with a wild ride in between), the model forecast house prices to rise 6.8% p.a. over those 2 decades. They rose 7.4%.
• Over the 00’s and 10’s combined, based on income growth of 4.1% p.a. and a 2.1% p.a, boost from rates falling from 7.27% to 3.28% (at end of 2019), the model forecast rates to rise 6.2% p.a.. They rose 6.4%.

So its not perfect, but pretty good.

Currently, the model says average house prices are about 10% above fair value. This may be a smaller number than what some would expect given the recent 20-30% spike in prices, but remember it reflects current, very low mortgage rates. And a 10% gap to fair value, positive or negative, isn’t an unusually large amount. Bigger variations to ‘fair value’ were seen in early 2008, 2010, 2012 and 2017 for example.

Market implied future interest rates

As I’m sure most are aware, the current RBA cash rate target is 0.1%. Based on longer term Australian government bond yields (e.g. 10yr is currently 2.79%), its possible to back out what is implied for the RBA cash rate at different points in time in the future. E.g. if currently the cash rate is zero, and the yield on 1yr bonds is 1%, that implies the cash rate in a year’s time is ~2%. Wow? 1% is the average return for the year, so to average 1% over the year, starting at 0%, implies finishing the year at 2%. Just some rough bond maths from a non-bond guy.

Currently, the market is forecasting the cash rate to be 1.2% in 6mths, 2.4% in a year, and 3.4% in 2yrs (source: Bloomberg). Two things to remember about this: 1) it doesn’t mean its forecasting the rate to be exactly 2.4% in a year, but rather a weighted average probability of it being either 2.25% or 2.5%, and 2) these are just implied rates based on currently available information. It doesn’t in any way guarantee that rates will be at those levels at those points in time. Think of them more like weather forecasts for some day in a month’s time. They are based on best current data, but we will get a lot more info between now and then that will move these forecasts around.

What happens to house price valuations under those interest rate scenarios

At the moment, the average mortgage rate for new lending according the ABS data is 2.5% (please don’t get caught up on the fact that you may get better or worse. This isn’t really that important to this analysis). So that is 2.4% above the RBA cash rate.

If we hold the spread consistent at 2.4%, a cash rate in 6mths of 1.25% equates to a mortgage rate of 3.65%. Using the data above we get a mortgage rate in 1yr of 4.8%, and in 2yr of 5.86%.

To return to a fair value that is falling due to these rate rises, but assuming 3% p.a. income growth, actual prices would have to fall 20% in 6mths, ~25% in 1yr and ~35% in 2yrs.

Some conclusions

• Distortions like changes in CGT, FHB grants etc, move markets around in the short to mid term, creating variations from the model. However, long term, income and interest rates are fairly reliable drivers of house prices.
• Those rate rises seen unsustainably high. Even if they take a bit longer to play out, they seem unlikely as consumption would collapse. I think its possible the RBA would blink before lifting rates that rapidly.
• I think a lot of people understate the impact interest rate declines over the last 3 decades have had on house prices increases. Without 7% p.a. income growth, 7% p.a. price rises aren’t normal for the average house.
• Prime locations may do better than that, but remember with all investments, high or low quality, entry points also matter. Buy in when the investment is over-priced, and returns can still be sub-par.

What does the data suggest for price growth?

1. Material price falls in the near term are really very possible, maybe even likely, unless we see inflation drop. However, the model is more about long term potential I guess.
2. Its very hard to justify material house price growth over the next 5-10 years from the current starting point under most scenarios.

• Although household income growth over the last 20yrs averaged 4.1% p.a., this was very front-loaded to the first 10 years due to a number of factors including better productivity growth, higher inflation and I believe the latter stages of the boost from 2nd income earners joining the workforce. In the 2010’s, income growth slowed and averaged 1.9% p.a.
• Its possible income growth is better over the next decade in nominal terms if inflation increases. But productivity growth is pretty poor these days (bar some covid distortions), and the double income household boost has mostly played out. RBA targets 3-4% wage growth p.a., so lets go with that for the expected contribution to house prices from household income growth.
• The other major variable in the model is interest rates. Unlike the major tailwind falling rates provided over the last 30yrs, with the cash rate starting at just 0.1%, at best we can say rates contribute 0% to house price growth over the next decade. Any increase from this level means a negative contribution to house prices.
• Very roughly speaking, a 1% increase in mortgage rates over a decade contributes about -1% p.a. to growth in house prices based on the model. In other words, a 3-4% increase in mortgage rates over a 10yr period could mean house prices go nowhere on average over that period. (note: this is just start to finish under this specific scenario. And even in this case, there would likely be a lot of volatility over the period.)

As severe as this sounds, its perhaps worth remembering that according to Christopher Joye, the RBA’s much more rigorous and detailed model suggests a 1% permanent increase in mortgage rates cuts prices more like 30%, so closer to 3% p.a. over a decade.

Anyway just some ideas. Appreciate any thoughts/feedback.

(Perhaps another conclusion from this is house price growth is all about broader inflation. But that’s maybe for another post.)