Property statistics can be misleading

by Alistair Helm in


When the news media report that the median house price in Auckland has risen by 13% over the past year, from $562,000 last March it's now $637,000 - that’s an increase of over $6,000 per month! 

This type of property news infers that property values are rising, and rising fast; as it's getting more expensive to buy houses - that would be a good interpretation - right?

Well actually no. You cannot make that inference. To hear that the median sale price is rising does not infer values are rising by the same amount. 

Now you may consider that I am arguing semantics. However I regard the difference between the rise in property value and the rise in median sales price highly significant. A significance that too often seems to be ignored or glazed over for the sake of a good headline.

The median price is a statistic calculated from the aggregated data of all the properties sales in a month in an area - a suburb, a region or the whole country. It should not be taken as an indication of the likely trend in value of a property in that self-same area.

Each year just less than 5% of all properties are sold - that is 1 in 20, which tends to amount to less than a handful of a properties on any given street around the country. The fact is that these properties recently sold might not be representative of all of the properties on the street in general.

For example, if I were to say to you that the median price for your street had doubled in a year from $300,000 to $600,000 would you think that the property market had gone mad and you were sitting on a gold mine? - sadly not. In this purely hypothetical example the likely fact was that the 3 houses sold last year were actually all apartments in that development at the end of the street and all the sales this year were of 4 bedroom properties along the street - you get to see the problem with the data?

This is in my opinion, part of the problem we have witnessed over the past few years in the inner city suburbs of Auckland. Suburbs that in some cases have seen rises of over 70% in 5 years. In these suburbs the properties being sold are in the main traditional 3 or 4 bedroom homes on 400 to 600m2 sections. So can we say here that values have risen by 70% in these suburbs as the median price indicates?

The truth is that rather as in the previous example, were the comparison was between houses and apartment, in these suburbs the comparison is more likely to be between un-renovated properties and renovated properties.

Look down any street in the inner city suburbs of Auckland and you will see the impact of gentrification and many hundred’s of thousands of dollars of renovation costs.

Here is a simplified summary of what has been going on. In 2009 and 2010 smart, savvy buyers (investors, developers and capable homeowners) started buying up ex-rental properties in these suburbs for prices around the $700,000 to $900,000 range thereby establishing the then median price for the majority of sales in these suburbs. Gradually over the past 4 years these properties have been renovated and placed back on the market following renovation projects costing anywhere between $200,000 and $600,000. The properties then have been selling for between $1,250,00 and $1,600,000 making a healthy return for those players in the market and in the end impacting the median sales price driving it up from the $800,000 range to $1,400,00 range a 70% rise. 

These renovated properties whilst in the main are still 3 or 4 bedroom properties on 400 to 600ms sections, however they are in many respects new homes as effectively all that remains of the original property is the floors, walls and roof. These improvements have fundamentally changed the inherent value of the property - as you would expect when a renovation costs averaging $400,000. But it is wrong, and does not follow that all properties in these streets / suburbs are now valued in the range of $1,400,000.

Here's a snapshot of inner city renovations of the past few years - then and now:

Further proof (if needed) of this situation, you need not look further than a couple of recent profile articles from the NZ Herald highlighting renovations in the inner city suburbs of Auckland


Grey Lynn - purchased Oct 2010 $604,000

Renovaton - new bathroom, new kitchen, redecoration

Sold 2014 $835,000




Ponsonby - purchased 2002 $620,000 

Current rating valuation (based on QV computer generated valuation model) 2011 $980,000

Extensive renovation

Auction 25 May - price expectation around $2,000,000

 


Now there is no doubt that these properties were rightly valued as a result of the selling price on the day. That was the price the winning buyer was prepared to pay for them, however does it mean that all of the other 80+% of properties in these suburbs have also risen in value by 70% ? - clearly not.

However as the saying goes and as all real estate agents are keen to observe - ‘A rising tide lifts all boats’. The problem illustrated here is that the statistic of median price is being misinterpreted as valuation.

The other compounding factor is that the source of property data on valuations, QV actually rely on sales price data to establish valuations. There computer based algorithm utilises recent sales data of similar properties (judged by size, not standard of refurbishment) to establish a current valuation. This valuation is then the benchmark by which they judge the sale price and that variance drives there monthly property vales trend analysis.

It is only a registered valuer undertaking a full assessment of a property through an on-site visit and evaluation that can accurately assess the true market value of a property.

 

Is there a better way? That is the question

Clearly a property sold after significant renovation or alteration cannot be judged to be representative of all properties in an area, whilst a property carefully maintained and resold after 3 years would be representative. Maybe this is the question that needs to be added to a more comprehensive data set collected by the Real Estate Institute. Through their comprehensive process of recording all sales by licensed real estate agents they can then provide a more accurate and representative price indicator so we avoid the inference that median sales prices are judged to be the same as property values, and in so doing improve the reporting of true values.

 

    

 


Making sense of the monthly property statistics

by Alistair Helm in


graphs iStock_000015752104Small.jpg.png

We are fortunate in NZ to be blessed with a rich and comprehensive database of property statistics. Many countries have to wait many months between data releases, which tend to provide quarterly trends. We have the benefit of insight into the property market with stunning regularity, all crammed into the first 10 days of each month.

However a common complaint is that the data is not consistent and it is presented using different metrics; for example median price, stratified median price, average price, valuation, seasonally adjusted truncated mean asking price, rental rates, to name but a few. So as a smart property buyer or seller it is important to understand what these metric can tell you and what you should pay attention to, to be ahead of the game.

There are 3 sets of data that I consider critical. These are property sales data (the measure of what is happening as a function of buyer activity), property listings data (the measure of what sellers are doing and how they judge the market) and then rental data (the measure of how landlords are pricing their rental properties).

Property Sales Data

This is probably the best-known sector of real estate data with both the Real Estate Institute (REINZ) and QV pitching their data to the market within days of each other. Data that on first impression appears similar, but on further investigation is somewhat different and the difference can be significant.

REINZ provides rich data on property sales, median prices and days-to-sell, across not just the major regions of the country but right down to clusters of suburbs. Sadly the full data set is not accessible online in a machine-readable format, rather it is published as a pdf with prior year and month comparisons.

The data is collated by the reported unconditional sales of properties by licensed real estate agents who are members of REINZ, submission (as I understand it) is not compulsory, but in spite of that it is very comprehensive, and its timeliness provides for the prior month insight within 10 days of the end of the month. The data goes right back to 1993 and using their online tools you can access databases for particular suburb clusters.

REINZ present the majority of pricing information in the form of median price, this is statistically appropriate as it ensures that extremes of sale prices within data sets don’t skew the data. However in preference to the raw median price I tend to focus instead on the Stratified Median price, this data set developed in cooperation with the Reserve Bank is sadly only published for the 3 major cities as well as the national figures. It is a far more accurate indicator of true price movements as it applies modeling to ensure that higher sales volumes in high price suburbs for example are normalized and thereby don’t result in an overall rise in prices.

QV produce a well-recognised set of statistics on the property market based on their rich database covering every house in NZ in their role as the government rating valuations organization. This database is updated on a daily basis by the transactions of settled sales as registered by LINZ. This process captures all property transactions irrespective of whether the transaction was undertaken by a licensed agent or a private sale (estimated at around 10% of all sales).

QV do not report selling prices, rather their business is valuation estimation and it is this index, which is published monthly. Their computational models analyse the actual sales prices for individual properties matched to prior valuations and thereby create an index of house price movements. This provides a good representation of trends of price movements rather than actual figures for property prices regionally or nationally.

One drawback to the QV data is that it uses a broad time period, each report is based on the prior 3 month’s settled transactions. This is further impacted by the fact that long settlement on some properties could mean the property sold unconditionally may not appear as part of the QV dataset for anything from 3 to 5 months. Despite this timing issue the trend indicators of price movements from QV are very useful and accurate.

Property Listings Data

Property listings data provides a vital insight into the supply side of the market and has only become available since the ascendency of the web as the definitive search process for buyers. The monthly data is provided by Realestate.co.nz in their monthly NZ Property Report published within a day or so of the start of every month.

The report details the number of listings coming onto the market in the prior month, the asking price of these listings as an indication of the sellers / sellers’ agent’s expectation, as well as the level of stock of houses on the market at the end of the past month. The report is very detailed in printed form and also provides the ability to download the full data sets with both raw data and seasonally adjusted data.

As the originator of this report during my time at the company, I believe the report  holds a unique insight into the supply side of the market, as a key lead indicator of the market. As an example of this is the fact that the current shortage of listings was flagged as early as April 2011 in the report by which time there was a clear trend as to how this would lead to price pressure in the medium to long term.

One key data set within the report that I think is worth focusing on is the Inventory as a measure of available stock of property on the market. This is presented not in absolute numbers but as a representation of the number of weeks that it would take (based in current rate of sales) in theory, sell all of the houses on the market. Nationally this has fallen from a high of over 52 weeks (a full years supply) to now barely half that at 27 weeks.

I have recently taken these key numbers and produced what I call a Property Dashboard - this simple gauge shows where each of the 19 regions of the country are at, in respect of experiencing a sellers market, a buyers market or a balanced market.

Rental Data

Weekly rental rates are published from the Department of Housing and Buildings (now known as The Ministry of Business, Innovation and Employment)

Tenancy Bond database. This monthly data is published in the NZ Property Investors magazine and provides great insight into the state of the rental market.

The Agency has recently opened up their data sets going back to 1993 by each local authority, so if you are keen to play with spreadsheets to analyse the data this is highly valuable.

Trade Me has also started providing some valuable data on the rental market. It is sadly rather infrequent, being on a quarterly basis and not as detailed as many would like, however I am sure that with the passage of time they will provide richer information, as they hold an incredibly rich data set of listings, rental prices and transaction pace, covering every property type, size and location.

Having begun by saying that we are fortunate in NZ to be blessed with a rich and comprehensive database of property statistics, I would actually conclude by saying that we are actually lacking real in-depth information and statistics. In a recent article I posed the question "Do we really have the property data we need?" I hypothesized as to what type of statistics would be really valuable, for example how would it be if we could for example understand:

  • What is the percentage of residential property buyers that are sold to first-time homebuyers, typically what are they buying, where and for how much; how has this changed over the years?
  • Equally imagine if we could understand how many properties in Auckland are bought as investment properties and how many of these are managed privately as opposed to being managed by a property manager?

Such rich data would provide so much more insight and assist consumers, economists and many other businesses to better plan and offer services.

This data is not beyond the bounds of capability. Real estate agents or the Real Estate Institute could capture all of this data in its capacity as the organization representing the industry and its professional practioneers.

This article is also published in the May edition of the NZ Property Investor magazine