Why RP Data publishes monthly price data

It was surprising to see an article published by Fairfax recently that criticized RP Data's publication of monthly housing market statistics. 

The article (that can be viewed here) remarks that the RP Data indices are ridiculous and implies that the monthly results are misleading.  The theme of the article focuses on the monthly result for Melbourne where we reported a 3.7% rise in Melbourne dwelling values over the month of July.

For starters, the article did not disclose the fact that Fairfax publishes a competing index which is not a monthly measurement and lacks the timeliness of the RP Data series.  Fairfax also owns a data business that is competitive with RP Data.  The lack of disclosure around this conflict should come as a shock to readers who expect a greater level of transparency in reporting.

It’s also worthwhile pointing out that our monthly hedonic index was developed (in conjunction with Rismark International) after the RBA released a discussion paper calling for more accurate and higher frequency reporting on housing market conditions. Since that time the RP Data monthly hedonic regression indices have become the index of choice for the RBA when it comes to reporting on housing market conditions in Australia.

In an attempt to discredit our reporting and highlight what they perceive to be extreme volatility in the index reading, Fairfax take the monthly movement in our index value for Melbourne and annualise it to suggest the Melbourne housing market is heading for a 44% plus rate of growth over the coming year.

Do commentators suggest the share market is set to boom when the S&P/ASX200 rises by 1.8% in a day (an annualised gain of more than 400% using Fairfax’s logic)? The graph below demonstrates the difference in volatility between our daily housing market index and the daily S&P/ASX 200 index.  The volatility in our capital city indices is around 4% per annum which is approximately one fifth of the volatility recorded across equity markets. The difference in volatility levels is clear from the graph below.

Housing markets are much more stable than equity markets, however that isn’t to say that housing markets don’t show natural volatility. Housing market volatility simply hasn’t been measured in the past due to the lack of granularity and sophistication in the methods behind the measurements.

If we look at monthly data released by the Australian Bureau of Statistics, we can also see some significant month to month fluctuations that can take commentators by surprise.

Over the first six months of this year, dwelling approvals have had the following month-by month changes: +9.0% in January, -6.6% in February, -4.5% in March, -6.1% in April, +10.3% in May and -5.0% in June.  Of a similar note, the number of owner occupier housing finance commitments has fluctuated significantly over the first five months of this year.  The fluctuations on a month to month basis were as follows: 0.0% in January, +1.6% in February, -1.3% in March, -0.5% in April and -1.7% in May.  The point here is that volatility is a part of measuring any market on a regular basis and must be considered when looking at the data.

Short-term movements in the property market cannot be viewed in isolation. A more reliable and informative way to interpret the statistics is to take the monthly change in context with the broader trend of the data. 

Over the past three months Melbourne values have shown a much lower 1.8% rise; in fact the rolling quarterly rate of growth has been tapering across Melbourne since March and the rolling half yearly rate of growth peaked in January.

Subscribers to our index data, such as the Reserve Bank and Federal Treasury, rely on our monthly indices regularly, but they take a much more informed view of the results; interpreting the monthly statistics in context with the trend as well as other market evidence such as vendor metrics, supply introductions, housing finance demand and transaction volumes.

Article continues on the next page. Please click below.

We agree that property market information needs to be timely and meaningful and it’s not just those people looking to buy or sell a home that should take interest. 

National regulators and policy makers have a high level of interest in housing market conditions; housing is an important component of financial stability.  The banking and finance sector track housing market conditions closely, with more than 65% of bank’s loan assets are secured by residential property their requirement to assess housing market conditions is one of the their highest priorities.  Industry professionals, who need to understand the ups and downs of housing market conditions on a timely and frequent basis in order to run a successful business, also have a requirement for accurate and timely housing market analytics.

Statistics are like a foreign language for most people.  While numbers can look straight forward, understanding what the numbers mean and how best to interpret them can be challenging. Unfortunately, when it comes to measuring things like asset values and market movements, whether it is real estate or equities or other asset classes, complex mathematics is often required.

Around 15 months ago RP Data, together with Rismark International, launched a new and improved hedonic methodology which provided a daily measurement of how dwelling values were performing.  Our daily index tracks the overall change in home values across the complete portfolio of properties across the capital cities; essentially imputing the value of every house and apartment each day and measuring the change across the overall portfolio values between each period.

Prior to being released to the market, RP Data’s indices were audited by both Alex Frino at Capital Markets CRC and KPMG from an infrastructure perspective.  We consulted with the RBA.  White papers on the index methodology are available to be downloaded from the RP Data web site.  No other index provider goes to such lengths to provide a transparent level of documentation around their methodologies and processes.

RP Data also spends about $15 million each year on data capture, cleansing and improvement.  On a weekly basis we typically capture between 85% and 90% of all auction results, which is generally much higher than other residential property data businesses.  We receive about 60% of our transaction data directly from the industry which provides a much more timely mechanism for collecting property data than waiting on government provided records, however we also collect virtually 100% of all property transactions nationally from the respective state government departments.

So in summary, we take data collection and the production and release of our indices very seriously as we fully understand and appreciate the important decisions being made every day by regulators, policy makers, financial institutions, other industry professionals and consumers, in reliance on our data and analytics, including our indices.

The nature of statistics is such that from time to time you will get outlier results such as last month’s Melbourne figure and yes, businesses with competing interests to RP Data will use results such as this to attack us from time to time.  The important thing is to understand outlier results occur every so often in any statistical model.  Like industry professionals, policy makers and regulators, industry commentators need to understand this, not dismiss or discount these figures and pay attention to the longer-term trends.