This snippet is from one of our previous episodes: The Biggest Danger People Face When Looking at Property Data.

We’ve all seen median house prices splashed across headlines… but how useful are they really for property investors?

In this TPC Gold bonus episode, Ben, Bryce and Suburbtrends founder Kent Lardner challenge the way most people interpret median prices—and explain why relying on that one number could send you down the wrong path.

Find out:

▶️ Why median prices might give you false confidence
▶️ How small sales volumes distort the data
▶️ The problem with beachside suburbs and micro-markets
▶️ What you should use instead for smarter analysis

You’ll also learn what Kent calls “the most consistent data set in Australia” and why zooming out from suburb-level stats could give you a more accurate picture of market trends.

Before You Buy, Get Real Data

Download our Free Property Report and see what’s really happening in your suburb.

__________________

If You Enjoyed TPC Gold | Are Median House Prices a Reliable Guide? You Might Also Like:

Transcript

Bryce Holdaway
So we want to pivot to getting your experience around analyzing property data and I guess the dangers of single data measures. So you’ve got a view on that. What’s the biggest danger people face when they look at stuff in isolation?

Kent Lardner
I think the first obvious one (is) the suburb median and a lot of people just look at the suburb median on its own, they don’t look at it through time. So the first thing you probably want to look at is the price distribution. We spoke earlier about the distribution of mortgages or the size of mortgages at a suburb level. Same thing, the distribution of price segments tells us a lot about the median; tells us a lot about that suburb. The beachside suburbs are going to be bimodal or trimodal; there’s going to be three or four markets in a beachside suburb. So it tells you a little bit about whether the median’s a robust measure or not.

The other thing I want to do is also look at how that’s changing through time. And the last thing you want to do is rely on just a 12-month change, because what you might find is things are dipping in the last few months. So if I only look at say: how has the suburb median changed in the last 12 months… great, it’s grown by 13% or 14%. But what it didn’t tell you is it’s dropped by 3% or 4% in the last couple of months. So you want to look at it as a continuous trend line through time. But you never really want to look at just the suburb, because the suburb is going to likely be volatile… volatile because it may have a very odd distribution of prices.

It could also be volatile because very few properties are selling. And that was our point earlier on. There are so many suburbs with very, very low levels of inventory, low levels of listing, low turnover. So I like to zoom out and look at the median at the SA3 level, a Statistical Area (Level) Three, kind of like the old local government areas, but a cluster of suburbs. So you get a lot more reliable metrics. So don’t just look at the suburb median on its own. Don’t just look at the suburb median and what it was 12 months ago. There are so many dimensions on a median price alone that you can really get an insight into a market from.

Bryce Holdaway
Kent, can you help the listeners really land that plane? So you talk at SA3 level, but you’ve got SA1, SA2, SA3 and SA4, right? So SA1 is concentrated within a suburb, SA2 is a couple of suburbs, SA3 is… can you just really land that plane for everyone?

Kent Lardner
Yeah. So an SA1 is the smallest area, give or take a couple of hundred homes that the Australian Bureau of Statistics… it’s the smallest area that they’ll divulge that census information without worrying about privacy. So effectively the SA1 then is like a Lego block that goes into an SA2. An SA2 is two to three suburbs big; just depends on where you are. If you’re in a really dense area, densely populated in a city… some SA2s are just one suburb large. But as you move into some of the regions etc., you might go four or five suburbs large. But on average, two to three is is pretty reasonable. What’s good about an SA2 is the naming convention. Usually they name it after the largest suburb in that cohort. I’m using the word cohort a lot, but that’s okay. You’ll forgive me. So effectively, you know, Lambton, New Lambton or whatnot. So it’s usually the largest most populated suburb goes first… so it becomes a little bit easier to understand.

And more often I often call an SA to a suburb in the media just to be friendly, because you know SA1s don’t have any names. They’ve got numbers. So you can’t really use those. SA3s… there’s a few hundred of those around the country. I typically analyze about 330 of them ongoing. Some of them align quite neatly to the old local government area (LGA). But what happened is Brisbane, for example, became the blob and it grew and it grew. And suddenly council amalgamations made it a huge city-size LGA. So reporting on Brisbane using the LGA is reporting on the whole city. But there’s so many submarkets in Brisbane; the East is different to the inner West. So carving up Brisbane using the SA3s was the catalyst for me to drop LGAs altogether and just go to the SA3.

The beautiful thing about the ABS and what it does in terms of carving up these areas, it’s designed specifically for statistical purposes. So if they’ve got to kind of draw a line and say, does that suburb go best there or there, they’ll align it to where it best fits statistically. So, you know, they’ll try and kind of smooth things out by tipping a suburb where it best belongs with other similar people, similar price properties, etc. And it works a treat. Whereas I kind of look at it: an LGA is designed for collecting garbages, and a postcode is designed for delivering postcards. So you know, the ABS stuff really works when it comes to property.

Ben Kingsley
Now, Kent, there’s a really big story in here right. Because for those people who are following property at home or are novices or don’t have this as their full-time gig… are you telling them that median house prices are unreliable to determine values of particular areas? Because I know that you and I have had great conversations about this because of exactly that point… that variability when it comes to different suburbs. And again, I mean, you can drive through different parts of a suburb and you can have a row of terrace houses, which are six meters wide. And then you have all of these other sort of Federation homes, Victorian homes where they’re 12 or 15 meters wide. But guess what, everyone? They’re all three-bedroom houses and the median house price of that three-bedroom house is X.

Now, I know I’m really basically putting a pin in the bloom of everyone who looks at medians and looks at performances against medians and they’re going… oh that’s not really great. And this is the challenge of analyzing property at all levels. We’ve got, you know, Kent in regards to SA3 levels; other people reporting on medians… inconsistency of median values across different data collectors. I mean, it’s just hard. And that’s why I’ve always said that anyone who thinks that you can just do everything from a data point of view without actually getting in the Suburban or doing flyovers using Google Earth and looking at the different natures of the land sizes, the topography and all of those other things… are effectively doing themselves an injustice. And that is the big danger for me in terms of what you’re saying about singling out of those particular data points as, okay, well, the median’s dropped by 15%. So that means that every house in that suburb is automatically worth 15% less. If you take that approach, you are wrong. You are blatantly wrong. And if you take an approach on long-term performance and so forth, you are wrong.

Now when we are providing advice and when we’re doing this type of work, we have to put some measure in there, right? So, you know, for most people, the preferred form of measure is median value for houses or for units. We don’t even have it at townhouses; is that an apartment, a unit, a villa? I mean, so what you’ve got to do is use that as a gauge and a guide only, and then you’ve got to start doing the legwork. You’ve got to start basically double clicking on all of those data points and then start to form an opinion. And that’s very hard across, as you say, 15,000 or 16,000 odd suburbs.

Kent Lardner
True. And there’s all sorts of scenarios where you’ve got say new stock being released. So pick on apartments for a moment. If you’ve got a new apartment block being released in that suburb, what impact does that have on apartment price medians? They surge, they go up. But actually the market risk can go up because you’ve got all this excess supply. It’s just that the median goes up because of the composition. I like to kind of look at a suburb median as just a descriptive stat. It’s just a description of what that data is… the midpoint of all the values lined up high to low at a point in time. And you come back a week later and that could be different again. So a couple of call outs is: what are you measuring in terms of your sample size? We know a suburb is a suburb, that’s okay… but are you looking at a 12 month rolling sample size? Or are you looking at a three month or a six month? And that’s probably the biggie. That’s probably the number one reason.

The other call out really is, you know, the areas that are volatile through time are going to be the ones that are more prone to having variances between providers… because that volatility through time is because I didn’t include that one and now I did, and that bounces around. So it could just be that not everyone’s captured that, that one particular property or those two extra properties. So often the giveaway will be, is there volatility in that line at that suburb level through that one provider or any of the data providers? Is it volatile through time? That is a good lead indicator to tell you it’s going to be prone to significant variances between the providers. And you’ve heard of hedonic index. Now I’m not here to talk about any particular brand, but hedonic index attempts to try and account for the variances in it and smooth things out and make it well-behaved.

And Ben to your example earlier on, if you had three houses that sold that were a very small lot size, and then you had a big house that was a 15 meter frontage… a hedonic approach would be to try and treat them all as the same size, make all of those adjustments. So they’re treated like exactly the same size property. So they’re all a 400 square meter property, three-bedroom, bla bla bla. So there are some techniques that you can attempt to overcome them if you attempt to measure the median, but there’s different ways to measure house prices. And I would argue that one of the better ways to do it is pull up a group of properties that you’re fairly comfortable with that represent the norm for the suburb. And then just do a retrospective appraisal of them as of a year ago. Group them together and that’s their price today. And go back and pull up those properties into a retrospective comparative market analysis or desktop appraisal as of a year ago. And by the time you get to the third or the fourth one, you got a pretty comfortable view of how that price has changed.