B UNDERSTANDING OUR STATISTICS


Calculating REINZ statistics, Median or Average?

Why the median is right for house price trends.

There are technical reasons why the median provides a more accurate picture of what is happening to the prices of houses rather than the “average.” As an example, assume there are 11 houses sold in a month with a price range of $200,000 to $300,000 and an average price of $250,000. Now replace one of those houses with a house that sold for $1 million. The average is now $318,182, even though 10 of the 11 houses for the month sold for less than this value. The median would be the price of the middle house sold in the range (in this case the sixth house), which more accurately reflects what the majority of the houses sold for.

REINZ uses medians to provide a more accurate measure of the “mid-point” of house prices that reflects what most people are going to be buying and selling houses for. 


Why Seasonally Adjust the Data?

Seasonal Adjustment is very common with economic and financial statistics. The reason REINZ seasonally adjusts is to account for well known and persistent effects in the data that mask what is really going on. The purpose of seasonal adjustment is to remove these effects and allow analysis of what the underlying trends are.

For example, we know that there are fewer house sales in December and January, and there are more sales in November and March. This happens every year. There is always a lift in sales in November compared to October; by seasonally adjusting the data we can see whether the lift in sales was just part of the normal lift in sales we would expect or whether something more important is going on. For example, in November 2011 the number of sales increased by 20% compared to October 2011. But after adjusting for the seasonal effect (sales always lift in November) sales increased by only 4.3%.

Thus 15.7% of the increase was due to the normal lift in sales we would expect in November and 4.3% was the increase after taking into account this normal lift . The seasonally adjusted figure is an important indicator of the underlying market trends. The timing of Easter can also affect house sales, for example if Easter is in March then sales may be lower compared to a March when Easter is in April. If there was no seasonal adjustment we might conclude that sales fell in the March with Easter for other reasons, rather than recognizing that the holiday may have lowered sales.

Seasonal adjustment, as the name suggests, takes account of effects that happen on a regular annual cycle. It can’t adjust for one-off events such as the Canterbury earthquakes or the Rugby World Cup.


Stratified Median House Price Index

What is the Stratified Median House Price Index and why is it useful?

The Stratified Median House Price Index is a more sophisticated measure of house price changes that takes into account differences between suburbs and locations. It was developed in conjunction with the Reserve Bank to better capture market trends. If, in a month, there were a larger number of sales in a higher value or lower value suburb than normal, the plain median is likely to be skewed to some degree. The Stratified Median House Price Index adjusts for any unusual change in sales mix to reveal underlying price trends.

To use the Stratified Median House Price Index to best advantage, take the percentage change between one period and another (month/quarter/year) rather than the level of the Index itself.  To see how it works, look at the line for New Zealand. It shows that the Stratified House Price Index for New Zealand has increased by 0.8 per cent for the last month, fallen by 0.7 per cent over the last three months and risen by 2.7% over the last 12 months. This indicates that prices have been moving up and down though the overall trend for the last 12 months has been for rising prices.

The “5 year CAGR” (compound average growth rate) shows the annual rate of growth (or decline) in house prices over the last five years. The “From Peak” column shows the price level of homes in the latest month from the peak prices reached in 2007. The table breaks this data down for the three main regions plus Other North Island and Other South Island and sections. So, for example, Christchurch prices are 1.9 per cent below their peak compared to NZ prices overall which are 3 per cent below their peak.


Need Your Own Statistics Report?

For bespoke statistics reports, please contact REINZ Business Analyst, David Shaw at dshaw@reinz.co.nz or Ph: (09) 359 5459. Please allow at least a few days as reports are not always available on the instant. Note that transaction level data is reserved for REINZ members only, custom reports for the public contain aggregate level statistics (e.g. medians, counts) only. Data is supplied in an electronic file. As a guide, small basic custom reports start at $250 + GST and increase in price dependant on the amount of data requested and the amount of time it will take to create the report.