For many people the appeal of property investing is quite simple. Aside from being less volatile than the share market, property is easy. Real estate markets move a lot more slowly than shares and are not characterised by the same speculation, emotion and wild price movements.

It’s a compelling argument – until you start looking at property numbers. Anyone familiar with these indicators will know that there are a lot percentage signs thrown into the equation. A plethora of data points seem to be in perpetual flux, while graph after graph shows a rollercoaster ride of ups and downs. It’s enough to make even the sturdiest of people dizzy.

With this in mind, we’ve decided this month to give you a complete low down of how the numerous data indicators affect the market. We’re covering everything you’ve always wanted to know about property data but have been too scared to ask: why property prices differ between data providers, how reliable property data is and how to use data to become a successful investor.

We’ve spoken to every data provider with a significant standing in the property research industry. Together they’ve explained how each of the data points they provide is compiled, what they reflect and how investors can use them to get the most accurate view of what is happening in the market.

Forget the dummies guides – you’re not one. This is the most comprehensive guide to how any investor, from a beginner right through to a seasoned buyer, can maximise their reading of indicators to score loads of cash, properties and maybe some bragging rights along the way.

A lucky country, indeed

To get a true appreciation of how useful real estate research data is, consider how the market was before many of these indicators existed. “Back in the 1980s, a lot of the information we now get was not being compiled,” says John Edwards, chief executive and founder of Residex. “It was not possible for ordinary Australians to get such a detailed view of how individual property markets were moving.”

According to Edwards, property data has helped to empower investors to the extent that they have access to information that is not attainable in most other countries. “Australia now gets a higher level of quality data, in a more timely way, than much of the rest of the world.

“The advantage we have is that each state has a centralised land title system and this makes it easier to access information. Data collection in the USA, for example, is much more difficult. In the UK, information has to be acquired at an almost a town council level, which is even more difficult.”

Edwards’ company is one of Australia’s largest compilers of residential real estate data. He founded it in 1990, initially to provide house price indices for the New South Wales government. The company changed this focus over the years and now has a database holding more than 350 million records of residential properties and sales data tracking back more than 150 years. 

Edwards says that property data is most useful for investors when it shows the general trend of markets. “Data can be subject to volatility on occasion and there can be anomalies, but what it does best is give you an overview of the underlying movement of a market.”


The premium source for finding out just how markets are moving is median property prices. By watching how much these prices rise and fall, investors can get a feel for how houses in a particular suburb are performing.

Andrew Wilson, senior economist at Australian Property Monitors, says the fact that prices growth is quite consistent over the long-term means it is possible to predict the growth potential of a market by how much it has over- or under-performed its long-term average.

“Studies show that since the 1930s, prices growth across locations usually aggregates to about 10-11% a year. We know this is able to remain consistent because of one primary driver: the aspirations of Australians to own their own home. It is this factor that keeps demand for housing up,” he says.

Wilson warns that this growth is not even. Between suburbs, towns and cities, there might be high growth in one area and low growth in another at any one time. Prices growth also tends to ebb and flow. Markets go through periods where price growth is high, sometimes even exceeding 10% a year, before hitting a ceiling. They grow to a point where they become unaffordable to the buyers they once appealed to, who now seek other, cheaper suburbs to buy property in. This results in a flat period of prices growth in one suburb and a return to price increases in another suburb.

How reliable are median prices?

While median prices give investors a great overview of property markets and make it easier to forecast when capital growth may occur, they are not gospel. They are subject to a number of anomalies that often come about because of the way in which they are compiled.

Australian Property Monitors has been releasing property price information since 1989, and Andrew Wilson says that even with this legacy of compiling information, his company and all the other research organisations face problems when putting together median prices.

“There are certain factors that make it an inexact science,” he says. “Not only is every seller in a market different, so is every buyer. We don’t know the conditions that a lot of sales were made in. Was the seller in a hurry? Was the buyer interested in other properties or just that property exclusively?

“Another constraint is that you’re using recent sales data to calculate prices growth. In a year you have only around 6% of all existing housing stock being transacted. If you drag that down to the quarterly level, this percentage is even lower and if you’re relying on this for your modelling, it means you’re only really getting a small snapshot of the market.”

RP Data’s senior research analyst Cameron Kusher says median prices also come into trouble when trying to separate the different types of property that are available in a market. “What one data provider considers to be a ‘house’ may be slightly different to that which another considers a house and the same goes for all other property types,” Kusher says.

RP Data maintains Australia’s largest property related database and Kusher adds that another problem that arises when putting together prices data is that there is no single methodology for calculating median prices. “Unfortunately, there is no standard Australia wide for the provision of property data,” he says. “Medians may be based on a different time frame and the process of cleaning, filtering and classifying the data may also be different.”

Refining the results

The different methodologies that each of the main data providers use all come about as a solution to some of the problems associated with putting together an accurate median price. Redwerks research director Jeremy Sheppard points out the biggest flaw – one that is accounted for in most methodologies.

“A simple measure of medians can get seriously out of whack,” Sheppard says. “Imagine a market where there are three sales in a month. Two sell for $200,000 and one, a premium property, sells for $400,000. The median would be the middle price, which would be $200,000. Now imagine next month there are three sales again. This time, however, two sales are premium properties for $400,000 and one is for $200,000. Taking the middle value, the median price would now be $400,000. Median prices would then show growth of 100%, even though none of the types of property saw value changes.”

Both APM and the ABS account for this discrepancy by using a stratified price modelling system, though the latter only measures houses, not units.

A stratified system separates capital city markets into price sectors that account for each suburb. In APM’s case, it divides the market into 10 price brackets, or deciles, according to suburb median prices. It then looks at how each decile moves within a period.

Wilson says this approach helps sort through some of the mess of raw data. “Rather than taking the whole market, we are taking sections of the market, so if there’s a particular bias at one end or the other, it is accounted for.” 

RP Data uses a different approach. It calculates prices according to a hedonic regression methodology. It may not be the easiest phrase to say with your mouth full, but Kusher says it has some key advantages. The technique uses recent sales data, combined with information about the attributes of individual properties such as the number of bedrooms and bathrooms, land area and the geographical context of the dwelling.

Kusher says deconstructing a property into its constituent parts allows for a reliable and accurate representation of capital movements. “[This way] the index takes into consideration those properties which have been transacted as well as those that haven’t.”

A further contradicting approach comes from Residex, which uses a repeat sales methodology. This arrives at a median for all properties in a suburb based on a calculated value for every house or unit. This allows recent sales activity movements to have some grounding in reality because a past progression of numbers is taken into account.

Who is right?

The short answer is no one. In fact, the advantage of having so many data providers working on different methodologies is that any time a statistical anomaly occurs in one provider’s information, you can often spot it by looking at data from competitor researchers. 

“You preferably want to look at as many stats as possible,” says Jeremy Sheppard. “Because the data providers all calculate things differently, occasionally one of them is going to be a bit off. If you see one provider showing prices growth of 6%, one of 7% and one of 19%, it is probably safe to assume the third, which is far off from the others, is probably a statistical anomaly.” 

RP Data’s Cameron Kusher believes there is space for all data providers. “The only risk of comparing data from other providers is getting a little confused as to why everyone has slightly different results,” he says.

Wilson adds that even when data points differ among providers, the information tends to point in the same direction anyway. “Most models do pick up, once all the data is in, the general trends in the market place. Different methodologies tend to produce different results over the shorter term, but over the longer term most models pick up the same type of underlying movements because they are checked against the same valuer general data.”