Land value matters, but how you assess land value matters more

22 April, 2024: Most property experts will be quick to tell you - land appreciates, buildings depreciate.

When we first developed our investment algorithms at FrontYa I wanted to test if this age-old saying was validated by data, or whether it was just over-simplified common sense.

Our research showed it was true, common sense triumphed again, although the common sense needed some tightening up. Land value can be one of the strongest and most consistent predictors of long term property growth, but it needs to be used in the right way.

The first problem is you never really know the land value. Sure, the State Government will estimate it for you, but given the land value determines council rates and land taxes, it’s reasonable to question whether it is a true reflection of the underlying land value, particularly as their valuation methods are appraisal based and can be materially different from how the market views it.

Nonetheless, they’re the only entity out there spending millions to estimate land values at scale for every property in the country, so to assess whether the land values drive property growth it’s probably your best data source.

The best way to think about land value is how much of the purchase price is derived from land vs. the building. The metric Land Value Contribution captures this as Land Value / Purchase Price.

In the chart below I’ve plotted how this metric stacks up as a predictor of house growth (I’ve removed apartments to control for property type, but we’ll explore those next). Each line represents the cohort of houses purchased in a given year, segmented by their Land Value Contribution. The y-axis is the average annualised growth of that cohort relative to the average for every house purchased in that year, e.g. a score of 1.7x means that cohort grew 1.7x faster than the average house purchased in that year.

Annual house growth (vs. median) by land value contribution

Source: Valuer General data. Analysis restricted to (1) purchase prices between 0.5x and 2.0x the average purchase price of the postcode and year to control for varying purchase prices, (2) properties held for at least two years to remove flips, (3) properties purchased and subsequently sold to enable an accurate calculation of annualised growth.

The chart shows the Land Value Contribution metric was great at predicting property market growth when you purchased in 2012 to 2014, but thereafter unless you’re buying at 70%+, the metric provided no useful information whatsoever.

The other problem is there appears a potential market cycle dependency here, which means you need to predict the market cycle in order to know if the metric will be useful. Given all these problems, our investment algorithms at FrontYa don’t pay too much attention to this metric on its own.

But don’t be disheartened! When massaged in the right way, land value is an excellent and consistent predictor of property growth.

The trick is to recognise that the Land Value Contribution is usually relative. The metric will typically explain why certain properties within an area grow faster than others, but it doesn’t usually explain why certain areas grow faster than others. This concept also explains why buying a plain plot of land in a growth corridor isn’t always a solid investment decision.

You can make the Land Value Contribution relative by controlling for location using quintiles, where quintile 5 represents the top 20% of properties with the highest Land Value Contribution for that area, and 1 represents the bottom 20%.

Annual house growth (vs. median) by land value quintile

The chart above shows that this relative Land Value Contribution metric does the trick. Not only can it differentiate fast from slow growing properties, but importantly, it is able to differentiate consistently throughout the market cycle.

This means if you can buy a house where it’s Land Value Contribution is in the top 20% for its area then chances are you’ll achieve an annual property growth rate of 1.5x the market. A worthy return indeed!

What about apartments?

In order to apply this principal to apartments you need to work out what percentage of the land parcel is allocated to the apartment. That is, if the whole block of units were sold, what proportion of the sale price will you get. You should be able to find this information in the strata’s title record, e.g. below:

Example of the unit entitlement on a strata.

This record shows that, for example, lot 1 is entitled to 8.1% (81/1000) of the strata, and so it owns 8.1% of the land value. You can then take the land value for the block and multiply that by 8.1% to get the land value attributable to your apartment.

When you crunch the numbers like this you’ll also find the above principal for houses generally applies to apartments too, albeit with less consistency.

The chart below shows the Land Value Contribution without the location adjustment:

And now with the location adjustment, to achieve a more consistent predictor.

This is great, but where can I find this data to help with my home purchase?

Land value data can be hard to track down, some States offer it free and available for download, others charge for access.

Good news for NSW and ACT is that it’s completely public and available for download from the Valuer General NSW or Access Canberra.

For everyone else, I would just try adopt the basic principal, and aim to buy older buildings sitting on great land parcels for that area. And extra points if that area has consistent buyer demand!

FrontYa is launching its Home Equity Investment Trust (HEIT), which provides investors with exposure to 2x the Australian residential property market growth. Click below to read more.