The great Auckland property divide

by Alistair Helm in


As the topic of affordable housing continues to capture the headlines; the statement is so often made that Auckland is one of the least affordable places to live in the world. Now I don't propose to debate this question, outside of making the point that if house prices were unaffordable, as in truly unaffordable; then the demand would shrivel up and price expectation would adjust accordingly. Now I know this is a simplistic statement, and the fact is that property prices have more than doubled over the past decade pricing many out of the market. However the fact remains that every day 60 properties are bought and sold across Auckland at a median price of $835,000. By comparison I see that the average advertised price for all property in Great London is now NZ$1,169,000 and in Sydney the sale median price is now NZ$1,225,000. The common link is that all 3 cities are centres of innovation and economic growth and are the leading city of their respective country. It is also not surprising that all three cities are simultaneously experiencing house prices falling, such is the connectedness of global property prices.

This current median price of Auckland property of $835,000 is 84% higher than the median property sales price outside of Auckland, where the latest median price is $455,000.

I was curious to see how this premium for Auckland property had tracked over the past couple of decades to see just how relatively unaffordable Auckland had become in domestic terms and thereby provide insight into the property price divide between Auckland and the rest of NZ.

I have chosen not to use median price as the metric for this analysis. As I have mentioned in preceding articles my favoured measure is the Stratified Price Index provided by the Real Estate Institute. This is in my opinion, the most accurate measure of house sales price for comparative purposes.

So this analysis is presented in the chart below, it is to my way of thinking quite enlightening. At this time based on the price index Auckland is trading at a 95% premium to rest of the country. Three years ago the divide between Auckland prices and the rest of NZ peaked at 128%, with Auckland more than double the price in the rest of the country.

The Auckland divide of house prices comparing Auckland with the rest of NZ

The historical perspective is very interesting. Back when data was first reported by the Real Estate Institute in 1992, property prices in Auckland were 40% above that of the rest of the country, through the mid 90's that shot up to 80% premium, a level that was pretty much sustained through until the end of 2013. Then began a surge in prices outside of Auckland, at a time when Auckland prices were certainly rising. This pushed the divide down to just under 50% at around the start of the GFC.

The next 7 years from 2008 Auckland went on a relative rollercoaster ride as compared to the rest of the country driving the divide from 50% up an ever accelerating steep curve to top out in 2015 at that level of 128%.

Given the current picture of property prices outside of Auckland compared to Auckland it is very likely that the divide will continue to fall in the coming year reducing the Auckland premium from 95% closer to 75% / 80%.


Do house prices quadruple every 20 years?

by Alistair Helm in


No sooner had I written the article "Do house prices double every 10 years?" than I get the question posted on LinkedIn as to whether house prices quadruple every 20 years?

My first reaction was - No Way, if they double in 10 years, how could they possibly quadruple in 20 years? - to be clear $100 doubling in 10 years becomes $200 and within 20 years quadrupling to $400. But then I remembered the good old adage of compound interest. The annual rate of growth to double a sum of money in 10 years is 7.2% and that same 7.2% annual growth rate compounded will quadruple the original sum within a 20 year timespan.

The next question for me was to Google the question, just as I did with doubling house prices, not 339 million results this time just 8 million. But top of the results from The Telegraph in the UK with a 2003 article stating just that fact. Between 1983 and 2003 average house prices went from £29,993 to £121,742 – a quadrupling. As an aside I bought my first house in the UK in 1984 for £22,150 so I can confirm that is what a typical house cost back then. I sold that house in 1985 for £20,650 so clearly I should have waited!

Anyway I digress. Back to the question posed on LinkedIn. Given the fact that NZ property prices have doubled during some 10 year periods, have we seen quadrupling over any 20 year period?

The answer is yes for Auckland but not for the rest of the country in aggregate. I have only been able to assess the period from 2012 as the available data from REINZ of sale price data only goes back to 1992. Here then are the same charts I used in the prior article adapted to track each 20 year period from 2012 to 2018.


NEW ZEALAND - ALL PROPERTIES

Analysis_of_NZ___Akl_pricing_since_1992_HPI___Median.png

So for the period from 2012 onwards the 20 year rise in house prices across all of NZ was an average of 237%, that is over a three fold increase but at no time has a quadrupling been experienced. Of note is the fact that the current period June 1998 to June 2018 at 271% is the highest ever 20 year rise.


AUCKLAND

Analysis_of_NZ___Akl_pricing_since_1992_HPI___Median.png

In the case of Auckland the average 20 year period price rise has been 296% between 2012 and 2018. In 31 of the 78 months the 20 year rise was more than quadrupling of prices in that 20 year period. The highest ever rise was for the period from August 1995 to August 2015 which saw a 329% increase.


NEW ZEALAND EXCLUDING AUCKLAND

Analysis_of_NZ___Akl_pricing_since_1992_HPI___Median.png

For the rest of NZ excluding Auckland the average 20 year price growth has been 197%, so just under a trebling of prices over a 20 year period. Of note though is the steady rise in the price rises since 2015 with the latest month of June recording a 240% increase in prices since June 1998.

 

 


Do house prices double every 10 years?

by Alistair Helm in


It's a fairly commonly held belief that house prices double every 10 years.

Is is correct? can it be guaranteed? and is there fact to back up the assertion? In researching this question, I found that asking the question on Google "Do house prices double every 10 years" provides a pretty comprehensive pedigree of search results, 339 million to be exact.

Among the top results I found that, an Australasian report by Core Logic which analysed the two 10 year periods from 1996 to 2006 and from 2006 to 2016. The earlier period did see consistent doubling (or more) of house prices across the major metros, whilst the later period only saw a doubling in Melbourne whilst Sydney managed a 78% rise.

The UK has similarly experienced doubling in house prices over 10 year periods with an excellent analysis by Alex King providing facts adjusted for inflation covering many decades.

Before diving into the NZ data, it is worth reminding ourselves of the mathematical fact that for a sum of money to double in 10 years requires an annual compound interest factor of just over 7%.


New Zealand - Total All Properties

Using the REINZ House Price Index as the measure of house sales price tracking back to 1992 I have created this chart below to best answer the question - do house prices double every 10 years? Simply viewed each bar represents the rise in prices for a 10 year period.

The black horizontal bar shows the 100% increase threshold. It can be very clearly seen that from October 2003 right through to November 2011, (with just a short period in 2009) property prices doubled, comparing a single month with the same month 10 years ago. The peak rise in any 10 year period was around mid 2007 (just prior to the GFC) when 10 year price rises topped out at just below 120%. 

The period since the end of 2011 though has refuted the claim of doubling prices every 10 years. Over the past 7 years, 10 year growth has been well shy of doubling, achieving around an average of 65% growth in each 10 year period.

Across the 26 years (198 months) the average 10 year rise was 88% with the median rise in any 10 year period 97%. Over that 26 year period just under half the time (43%) saw prices doubling or more


Auckland

So what about Auckland, the leading metropolitan region of the country? - most of the media portray the Auckland market are constantly bubbling and among the most expensive cities to live in. What does the data show?

This chart is certainly illuminating and somewhat surprising. Compared to all of NZ, Auckland has sustained a relatively consistent higher 10 year growth in prices with an average 10 year growth of 98%, well ahead of the NZ total at 88%. However the periods during which a doubling was seen were fewer and not aligned to those of those of total NZ. In just 72 of the 198 months was a 10 year doubling witnessed - less than the 85 for total NZ and the highest 10 year rises where in early years of the new millennium and the recent boom cycle.


Total New Zealand excluding Auckland

If you extract the Auckland price rises from New Zealand in total you get a very different answer to the question - do house price double every 10 years?

During the period of 2005 and 2008 in the run up to the GFC the 10 year respective rise 10 years later as seen between 2015 and 2018 property sales prices outside of Auckland would be barely 40% which is an equivalent rate of 3.4%, certainly ahead of inflation but probably below the prevailing mortgage rates for the period.

So can it be said that property prices double every 10 years? - clearly not; as to make such a statement is always made assuming the past is the best indicator of the future which we all know is wrong. However it would be true to state that in the case of Auckland that over the past 20 years property prices have pretty much consistently doubled every 10 years as the median 10 year rise is 96%. 


Subsequent to writing this article I was asked via a question on LinkedIn as to whether property prices quadrupled over a 20 year period. I had not heard such an assertion, however upon Googling I found it was quoted as a fact. So I naturally went and did the analysis and then wrote this article - Do house prices quadruple every 20 years?


The property market cycles of the past 25 years

by Alistair Helm in


When property prices are on the rise, the commonly asked  questions are usually based around how long can this bull run last?; are we in a bubble?; or have prices become disconnected from reality?

All are fair questions. However what is really interesting is the fact that these questions tend to float to the surface of the media precisely at the end of a property cycle. Property cycles are a historic part of the property market, here in NZ and globally. These cycles tend to last around 5 years and represent periods of either stagnant growth, accelerated growth and in some case declines.

In an article I wrote a few months ago I isolated the key periods of the property market identifying 5 periods tracking the median price data provided by The Real Estate Institute for over 25 years. Those 5 periods comprised 3 periods of rising prices and 2 of plateauing prices. The chart below reflects this perspective.

Interestingly I was sharing this insight the other day to a group of real estate colleagues as we reviewed the state of the current market. We are currently in a plateau phase - certainly in Auckland. The question I was asked of this chart, was as to the relative rate of increase during each of these periods as there is a perspective that property prices grew faster in the recent 5 years as compared to the 2002 to 2007 period and certainly as compared to the 1992 to 1997 period. I pointed out that this chart can be deceptive. The representation with a vertical axis of the actual median price can be misleading as a 100% increase from $100,000 to $200,000 as was the case in the 90's represented a dollar rise of $100,000; however a dollar rise of $100,000 in say 2008 from $350,000 to $450,000 only represents a 29% increase.

This got me thinking as to whether there was a better way to represent the trends of these cycles.

A couple of hours work has brought me to this new analysis of the property market cycles of the past 25 years.

Firstly I have chosen to use a different data set from The Real Estate Institute of New Zealand for analysing the market. I have chosen to use the House Price Index (HPI) - this data set is probably the most reliable and accurate data set for property sales prices in NZ. The Index which was developed by REINZ in cooperation with the Reserve Bank of NZ ensures that this price data set is not effected, as it can be by composition changes. That is to say if in a month a higher proportion of sales occur in a relative price band or suburb set this can skew the median price which is the most common and trusted method of sale price calculation.

I also decided that I would analyse two data sets namely Auckland, and NZ excluding Auckland. The fact is a single NZ data set is itself a compositional skewed data set as Auckland being c.40% of the market heavily skews the National picture which then becomes irrelevant for properties outside of Auckland. (More of this is a future article).

The chart below tracks the House Price Index from 1992 to date, for the data set of Auckland and Total NZ excluding Auckland. The Index is centred in 2003 at 1,000 points with dates before that being below 1,000 and dates afterwards being above 1,000. A figure of 1,100 represents a 10% increase over the base month in 2003. A figure of 2,000 being a 100% increase over the base month in 2003 and so on. The latest HPI data from REINZ for May 2018 had an Index of 2,883 for Auckland and 2,554 for all of NZ excluding Auckland.

This chart is very illuminating as it shows very clearly the plateau the Auckland market has been experiencing since reaching 2,800 in August 2016 - nearly two years ago. It also very clearly shows the extent to which the Index for NZ property sales price excluding Auckland shot away in the 2003 to 2007 period. The index to 2003 though in someways confuses the visualisation. I therefore decided to recalibrate the index it back to 1992, as the revised chart shows below.

Now back to that opening question - what are the key cycles of the property market of the past 25 years and how similar or different are they?

Here is a more detailed analysis of the price index data I have concluded upon for which I have identified not 5 but actually 7 phases of the property market over the past 25 years.

Now before you email or call me to ask what on earth I have created, let me explain. I have taken the HPI data and broken it up into these 7 periods or phases of the past 25 years which are tracked along the horizontal axis from 1992 to 2018. For each of the 7 phases I have identified the inflection point and recalibrated the start point to zero. So let me walk you through the 7 periods to help explain.

Period 1 - January 1992 to February 1993

Just a one year period which saw the price index grow by just 7% for Auckland and 7% outside of Auckland. This is subjectively an odd data set as there is no record of what happened pre-1992 and especially from the after affects of the stock market crash of 1987. Best to disregard this phase.

Period 2 - March 1993 to March 1996

Now this was a period of significant price growth. Auckland price index leapt 68% in just 3 years whilst outside of Auckland the price index went up 22%. This 3 year rise of 68% in Auckland is the fastest rise that has ever been seen over the past 25 years.

Period 3 - April 1996 to January 2002

This period of close to 6 years included the Asian Financial crisis and saw property prices stagnate. By the end of the period Auckland price index had risen by just 10% and the rest of NZ excluding Auckland 12%. Twice during that period Auckland prices dipped. By 5% in the first few months after April 1996 and then again late in 1998 dipping by 3%.

Period 4 - February 2002 to November 2007

This would be viewed by many as the most explosive period of property price growth. In the five and a half year period the property price index for NZ excluding Auckland more than doubled, rising 116%, far outpacing the rise in Auckland of 89%. All of this of course foreshadowed the GFC.

Period 5 - December 2007 to December 2009

For a full 2 year period the property market went into a funk. Prices went down. A scenario that many thought would never happen just a few years earlier, however property prices will at some point, just as with all assets; go down in value as much as they may go up. For Auckland the price index slipped 12% in the first 12 months before easing back to be down 3% by the end of the second 12 month period. For outside of Auckland the level of price fall was less severe, down 9% in the first year before clawing back half of that to be 5% down after 2 years.

Period 6 - January 2010 to January 2013

A full 3 years in which property price index rose just 2% outside of Auckland, spending the first 2 years underwater adding to the prior 2 year period making a full 5 years when properties collectively outside of Auckland failed to recover the peak they had attained in late 2007. For Auckland the recovery from the GFC came earlier with this 3 year period seeing a modest sustained growth up 18% over the 3 years.

Period 7 - February 2013 to date (May 2018)

This current 5 year period may in later years be revised to a 4 year period ending in August 2016 when Auckland's property price index peaked up 70% in that period of three and a half years, close on the same rate of growth seen back in the '93 to '96 period. At the same time for properties outside of Auckland the price index keeps rising, up 49% over this 5 year period, the second highest rate of growth of the past 25 years.

So in conclusion, property markets move in cycles and these cycles lead to price movements. It is likely, based on historical evidence of the past 25 years that we are in, or will be shortly be entering a plateau phase, typically these periods can last from 3 to 6 years dependent upon the background economic circumstances. As to when this period started, in the case of Auckland we may already be 2 years into it. In the case of the rest of NZ it maybe that we are about to see this plateauing start very soon. If there is any silver lining it should be seen that a correction that sees property prices fall is less likely and with a strong economy we are likely to see property prices take off again within the next decade. The only certainty is that it is almost always impossible to identify the infection point until after the event !


Are we really facing a “serious shortage of properties for keen buyers”?

by Alistair Helm in


Shortage iStock_000009057106Small.jpg

It's funny sometimes when you look at a piece of research which on the face of it looks to be enlightening, providing as it does new facts and insight; but when dig a little deeper you discover that it merely reinforces known facts.

Such was the situation this week when Realestate.co.nz published a research study by Horizon Research of consumer expectations of buying and selling in the coming year. I was at first somewhat confused, as the report was completely buried in the monthly NZ Property Report. I was not sure that they had clearly thought through how they were presenting the data, however by tuning out the monthly stat charts the research data findings seem at first eye-opening.

The research results show that 6.4% of the adults 18 years or over surveyed are definitely looking to buy in the next 12 months, but only 2.8% of people stated they are definitely looking to sell in that same period, says realestate.co.nz spokesperson Vanessa Taylor.  

This equates to a shortage of around 3.6% of homes across the New Zealand marketplace, representing a significant number of properties currently not on the market, she says.

Even taking into account new builds which are underway and not included in these numbers, there will be an increasingly serious shortage of properties for keen buyers.
— Vanessa Taylor - Realestate.co.nz

Wow – 6.4% of adults definitely going to buy; but only 2.8% going to sell! As stated that would seem to be a “serious shortage of properties for keen buyers”

The danger as ever with statistics is that seeing one set of data in isolation can lead to misinterpretation. Sure, in this survey 6.4% of adults are going to buy vs 2.8% are going to sell. On the face of it, it certainly seems serious and due cause for alarm, especially when we are constantly being told in the media (and by politicians) that we need to build more houses to solve a housing crisis. But hold on a moment. Step back and ask what is the context of this research data. We need to ask how have results changed over the past year or the past 5 years? Are we seeing a growing trend or a declining trend? After all one data point does not tell us everything.

The beauty of the internet is that you don’t have to hound people to get them to provide the data, the data is accessible. Horizon Research has been undertaking a Housing Supply and Demand Survey since 2010 on a fairly frequent basis and here are the results.

So it would appear that the data for May whilst perceived to be high was actually down on the previous report from October last year, when the level of definite buyer activity was 9.8% and seller intention was just 2.5%. That’s 1 in 10 adults saying that they were going to definitely buy. At the same time just 1 in 40 adults said they were going to definitely sell!

How did I possibly miss that piece of news?

There is no denying that the chart very clearly shows an ever growing divergence between buyer intent and seller intent, however the question has to be asked. If as the media release this week states … will there bean “increasingly serious shortage of properties for keen buyers”?

How could this be true? Especially if we have witnessed this divergence for the past couple of years. If it were as true as the statement leads us to believe, then surely we would be seeing rampant price inflation as this pressure on supply would seem to portend?

To answer this key question we need to separate the buyer and seller data in the research.

The Buyers

First let’s deal with the data of buyer intent. The latest May research states that 6.4% of adults definitely intend to buy in the next 12 months. Let's work through the numbers. In NZ there are 2.7 million adults; being 60% of the population by age between 18 and 65.

The research states that the findings based on a sample size of 1,345 adults from a nationally representative panel show that 6.4% state that they are definitely going to buy a property in the next 12 months. The average household composition of adults in NZ is 2.02 adults per household which equates to 1.34 million properties. This reflects the 85% of private property occupied by adults under 65 years. If you apply the 6.4% proportion from the May research to this total of 1.34 million properties you arrive at the figure of 85,700 properties definitely going to be bought in the next 12 months.

A total of 85,700 properties being bought in the coming year is in fact a perfect estimate of the market. In the past 12 months volume sales as reported by the Real Estate Institute were 74,600. Sales volumes have begun rising from a low of 73,500 in the 12 months to December last year. So I would say that the research is bang on as an estimator of market demand, providing an accurate guide as to the likely pick up in sales volumes for total 2018/19.

What is also interesting in extending this calculation; is that the historical research carried out by Horizon Research shows that the past 8 years data correlates quite closely (with just a few outliers) with the market trend of moving annual total sales as highlighted in the chart below:

The Sellers

So how can the divergence of buying intention from selling intention be explained? How can it be that what we can validate that 6.4% of adults definitely intent to buy, but clearly far more than 2.8% definitely intend to sell? In fact we can I judge with confidence say that 6.4% of adults will sell in the next 12 months.

Herein lies the answer, in my opinion.

Existing homeowners go through a process. A process that ends up with the purchase of one house and the sale of another, however this process from a psychological perspective does not begin with a rational intent to sell. It starts with an intention to buy.

Selling a property is a means to an end; the end being the next home. It is a functional process, not an emotional decision process – the emotions lie in the expectation around the next home. So when asked in a survey the intention to sell, the typical adult probably under-reports, largely due to the anxiety and expectation surrounding the process of selling. That is my opinion.

So I don’t support the view that we are really facing a “serious shortage of properties for keen buyers”, we are simply seeing the normal property market at work.

This thinking is very enlightening to the processes of real estate industry in general and particularly to the marketing strategies adopted by players within the industry. I highlighted recently in my article about the likely launch of Better Homes and Gardens Real Estate how their brand positioning would not play out so well in NZ as the market here is a vendor (sellers’) market and their brand is all about emotional inspiration and lifestyle.

The fact is NZ real estate companies largely focus on the processes of selling and buying linking the brand to the "For Sale" event, but I have not seen many (if any) real estate companies, nor for that matter any real estate websites position their brand around the emotionally engaging process of discovery, in the process of finding your next home.

 

 

 

 


The Auckland quarterly property review - Q1 2018

by Alistair Helm in


Quarterly Review logo.png

Having summarised the broad New Zealand property market for the 1st quarter of 2018, it is critical to examine separately the Auckland property market.

Auckland is a very different market to the rest of the country by fact of scale and greater international influences. Additionally Auckland tends to be a bell-weather to the broader regional market, thereby investigating the local market trends provides insight both for Aucklanders and as a future indicator for the rest of the country.

As far as volume sales are concerned it appears from the latest March data to the end of this first quarter of the year, that we seem to have bottomed out. Sales of properties have been falling for close to two and a half years. Seen on a 12-month-moving-total the Auckland market peaked in October of 2015 and has fallen in volume terms consecutively by 37% since then to the current March 12-month-total of 21,350 sales. We do seem to have avoided dropping below the 20,000 sales a year threshold experienced through the GFC and the rebound in 2011.

The chart below tracks the monthly variance for a year-on-year comparison of Auckland property sales for the past 18 years. It is clear looking at the chart that the market experiences significant volatility in sales movements in Auckland, up over 50% year-on-year at times and equally falling by similar variances. Since that peak in October 2015 the variance has been consistently negative with just couple of months where there was a small correction. As noted in the wider NZ analysis the March month this year did see a surprising fall year-on-year in sales but this is not unusual as can be seen in prior market cycles of the past 18 years.

As I have mentioned many time in the past in the context of property market commentary, a critical issue in NZ analysis of property sales is the number of dwellings and how that has grown over time. In the case of Auckland, hardly a day goes by when the media does not refer to the 'shortfall in housing' affecting the city - whether that shortfall is 20,000 or 50,000 the fact is Auckland has grown at a staggering rate over the past 25 years. 

Based on the trending of the last census data it is likely that Auckland now has surpassed 500,000 dwellings - this is up from around 445,000 10 years ago. This growing level of new dwellings naturally will be a factor in assessing the true level of property sales. Tracking this over the past 10 years further reinforces the market view that we are bottoming out of the cycle at a low level of 4.3% of all homes being sold in the past 12 months, this compared to a 10 year average of 5.4%. The broader NZ position interestingly is that 4.5% of all homes were sold in the past 12 months as compared to a 10 year average of 4.7% emphasing that the Auckland market has fallen in volume terms further than the rest of the country.

As I have often stated I am of the belief that watching closely the sales volume trend is a better indicator of the state of the property market than following the median price, as price is largely a reflection of the state of the market rather than an indicator. This is best demonstrated by the chart below. This analysis which I introduced a couple of months ago tracks the clearance rate to the median price movement. Clearance rate is the relationship between the new listings coming onto the market in a 12 month period of and number of sales.

This latest update to the Auckland chart of Clearance rate to median price shows again evidences the bottoming out in the clearance rate and the start of some degree of increase in the past 4 months, whilst at the same time the median price variance year-on-year is showing an arresting of the fall seen in the past 2 years with the current situation seeing median price level or slightly down compared to this time last year.


The NZ quarterly property review - Q1 2018

by Alistair Helm in ,


Quarterly Review logo.png

The first quarter of 2018 is now behind us and with it the final month of the financial year. It is interesting as an aside that this last month of the financial year, through no direct result of any promotion or marketing by this industry towards the consumer, typically ends up being one of the biggest sales months of the year. Sure it's influenced by seasonal considerations but their is no denying the fact that incentives within the industry do bring more properties 'across the line' into March than would be the case without these incentives.

With that as a backdrop it is a good stage of the year to examine the state of the market from the available published data from both Realestate.co.nz and REINZ.

I always like to start any property market analysis with sales volumes as this for me is the most important driver. The Median Sales price whilst the headline of all property news articles is largely an outcome of the market, the result of the pressures of the market; but the number of transactions indicates the real health of the market.

Across the whole of the country, sales volumes have been weak now for what is close on 2 years. As the chart below shows, the market year-on-year variance dropped into negative territory nearly 2 years ago in July 2016. 

In the middle of last year after 12 months of declining sales we started to see the rate of decline ease off, and as we reached Christmas, year-on-year sales were pretty close to level with prior year. This was followed in January and February of this new year by very slight rises year-on-year, but then March seems to have hit us and we are back year-on-year to a 9% decline. This is why I made the comment earlier. I had expected that March would have been stronger than it was. For whilst comparing the same influence on the market last year, 9% decline after what has been very low sales volumes is surprising and has that feel of a 'late spring frost' arresting the early new spring growth.

Yet despite this 'snap of frost', it is my view that we are still likely to see sales volumes rising through 2018 to end well ahead of 2017. The most recent 12 month total sales volume to March was exactly 73,000 as reported by REINZ. This level of sales is almost identical to the low point of the last cycle in October 2014 before the upswing in sales that lasted for the next 18 month as sales rose to a 12 month moving total of 94,000.

When examining property sales over a long term perspective, a key influence that has to be factored into the analysis is the core underlying growth in the number of properties across the country. Twenty five years ago back in 1992 when REINZ started collecting and publishing real estate statistics there were just under 1.2 million properties across the country and in that year total sales amounted to 63,000. At the end of 2017 there were an estimated 1.63m properties, an increase of over 435,000 properties, up 36%, whilst total sales in the 2017 calendar year was 73,557, an increase of less than half that of the growth of number of properties. Property sales have gone through around 8 cycles in those 25 years reaching a all time high of 121,777 in April 2004 and an all time low of 53,463 in February 2009.

These highs and lows of the market have represented, a proportion of sales to actual dwellings with a high of 8.4% and a low of 3.5%. Across the past 25 years the average has been 5.7%. Over just the past 10 years the average has been around lower at 4.7% with the current level of the 12 months to March 2018 at 4.5%, certainly below the 10 year average and the longer term average. This further reinforces the view that the market is lower than would be expected.

Complementing the sales component of the market assessment, I am keen to examine the latest data of the clearance rate. This metric I introduced back in January when assessing the year-end data for 2017. It is the measure of sales as a % of listings applied to the latest 12 month moving total.

The picture for clearance rate which as I demonstrated back in January tends to track pretty consistently to median price movement over the years. Looking at the most recent 3 months the clearance rate appears to have arrested its decline and similarly the annual median price movement has stablised at around 5% allowing for the monthly volatility.

Taking all these data points into consideration it looks to me that we have reached the bottom of a cycle of property sales. Given the scale of the current residential dwellings at a level of 1.63m, a sales level of 70,000 is a low point and over a forthcoming period of what maybe 2 to 3 years we will likely see sales rise up again to an expected level of around 90,000. This assumption is predicated on the belief that a sufficient flow of new listings will come onto the market to facilitate this lift in sales, for without this, the latent demand will be unsatisfied and that has the potential to stall the market. 

 


So.. should I wait to list in the Spring?

by Alistair Helm in


4 Seasons trees iStock_000007235396XSmall.jpg

Sitting in a recent sales meeting of our local office the other day, the discussion came around to the pipeline of forthcoming listings. The one consistent comment emerging from around the table was the client desire to "wait for Spring". I have long heard this view that Spring is the best time to sell, but is it really?

Spring may well be the busiest time for new listings, but is such a cluttered and contested market the best time to grab and fight for attention?

This question and the data to support the opposite view is something I have written about in the past, but I thought it was about time to revisit this matter for the benefit of buyers, sellers and my new colleagues in this real estate industry.

From a statistical point of view the way to assess the seasonality of the property market is to create a picture of weighted average sales for each month, based on the historical sales by month. This is the approach I have taken in the past. However when updating my spreadsheet it did strike me that this approach was somewhat "rough & ready". After all, months have different number of days, especially when you allow for the major public holidays. So I factored this in to created a weighted average sales per day for each month and then expressed this in the chart below as to the variance that each month represents as compared to a "normal month".

This charts shows that the most active month of the year in terms of sales is March with January the least active. October is the month which could best be described as 'average'. The overall trend is that the most active sales period is February to May. The winter months through to September are quiet and then November comes storming back before December sees sales tail off. Nothing surprising in this I would have to say. As a note this data is total NZ sales from the full data set of Real Estate Institute statistics from 1992 to 2017.

In doing this statistical analysis, as very often happens I questioned the data to see if once you broke up the data for the full 25 years into separate 5 year periods, if this pattern has changed in any significant way?

This then is the same analysis but showing each of the five 5 year periods from 1992 to 2016.

This analysis I find very interesting (but maybe I am the only one!) - the key months for me are January, May, July and December as these to my mind are demonstrating a valid statistical trend.

January is very clearly becoming a much quieter month. In the most recent 5 year period (2012-216), a period in which sales have been buoyant, it has been close to 30% quieter than a typical average month. Why? Well it could be a factor of lifestyle and people really wanting to enjoy summer and not worry about buying or selling a property. However I believe the real estate industry has begun to see the need to manage their business better, and so establish campaigns that culminate with sales pre-Christmas or wait until the end of January and in so doing leave January clear for vacation. 

The month of May is very clearly becoming a more active month with a successive shift from being the 5th quietest month back in the early 90's to being the 3rd most active month of the year in most recent times. Why? I suspect that the earlier statement about how January has become less active has shifted the summer marketing campaign period well into May.

July has like May, become a more active month, however it is still the 3rd quietest month of the year.

Finally December has become less quiet, this coupled with a slight lessening of the activity in October would possibly indicate that the traditional Spring market activity is getting later and the lead up to Christmas is as active as the main month of November, especially as it lends itself a settlement and house move in the best of the summer days in the new year. 

 

Back to the matter of is Spring the best time to sell?

 

To really assess the best time to sell, you need to look at both the listings and sales by month to picture the seasonality. Statistics for listings only go back to 2007 via Realestate.co.nz and their NZ Property Report, so I have adapted the seasonality of sales chart to measure the past 11 years. So here are the charts based on this 11 year period for sales and listings.

These two charts have the same identical axis scale of -25% to +25% and the first thing that strikes you is how much more volatile listings are in terms of seasonality. There are only 5 months of the year when listings are more than average (Feb/Mar/Apr/Oct/Nov) with 7 below average, whereas sales are split 50/50. For sales there are only the two months of January and March when the variance to the average is greater than 10%, whereas for listings there are only 3 months when the variance is less than 10%.

A better way to view this mirroring of supply and demand is to line up the two sets of data on a single chart.

This is the chart that best shows the challenge of when to list a property for sale. If you hold by the adage of being a 'big fish in a small pond' then the best time to list is during the winter when listings drop off far more than sales drop off.

If you wanted to call a single month then December would be the winner. The differential between listings and sales is 16 percentage points, closely followed by May at 13 percentage points.

The worst month to list based on this analysis is October where the differential is 17 percentage points. This would therefore seem to completely quash the notion that the best time to sell is the Spring, certainly a deluge of new listings hit the market in the Spring but based on the latest 11 years of NZ sales data, September and October are actually quiet sals months with November being the single active month of the Spring.


Online Property Valuation Models – how accurate are they?

by Alistair Helm in


iStock_000004287274Small.jpg

As might have been anticipated, my recent article providing a guide to the current portfolio of providers of online property valuations models triggered the inevitable question – "just how accurate are they?"

So I thought I would do some desk research. However before I unleash a barrage of criticism stating that there are heaps of examples where the Automated Valuation Models (AVM’s) are so wide of the mark to make them laughable, let me simply say this. There over 1.5 million AVM’s or potential AVM's for NZ properties – there will always be outliers and extremes. I do not have time nor patience to review thousands of properties, or even hundreds of properties. I chose to select just 12 properties.

The method I have used, is to track the latest auction results as published by the team at Interest.co.nz as the auction year started after Christmas. I simply took the first 12 I saw which comprised 8 properties in Auckland and 4 in Tauranga. So again I acknowledge that my sample is hardly representative nor truly random. It is made up of auction sales only, the sales are only for those 2 areas of the country and represented a very quiet period of the year.

With these 12 property sales results I went to each of the 5 providers:

I knew none of these providers had updated their valuations to take account of any of these actual 12 sales neither would the sale records have been picked up through local council sales or agent reporting so there was no bias of an AVM being influenced by these recent sales.

Another point to note is the analysis compared the sale price at auction to the mid-point of the price range of the AVM.

So here is the table of results. The colour code used is blue where the AVM equalled the sale price exactly, red signifies an AVM below the sale price with green where the AVM is above sale price. Finally, grey indicates that the provider had no AVM for the property.

As you can see, the visual skew towards red indicates that based on this sample set most AVM’s were below sale price.

The original version of this article I used an average variance measure, after receiving valuable feedback I have now used the calculation of Gross Median Error.

All providers achieved a gross median error of less than 10%, with Realestate.co.nz achieving less than 5% which is impressive. I would deduce that a factor in their accuracy, is they benefit from the very latest REINZ data each month of unconditional sales, whilst all other provides rely largely on settled sales which come through at least a month to 2 months later.

Another perspective I was keen to examine in respect of the accuracy of AVM's was the indicative range they provide to reflect the level of confidence. For each provider, for each property I assessed the range as a percentage of the midpoint price.

This analysis is very illuminating. The provider with the tightest range (in theory indicating confidence factor) is MyValocity, closely followed by Homes, both just under 10%. This effectively meaning that their AVM range is 5% below the midpoint to 5% above which I would judge as fairly acceptable given this is a computer based estimation with no detailed knowledge of the specifics of the property.

Of interest in this analysis is the very wide margin in the range from Trade Me Property at close on 30% with their tightest range being for a single property at just 19%. Similarly Realestate.co.nz seem to apply a standard c.21% to all AVM’s.


For completeness here are the raw numbers

 


Asking prices and selling prices - a comparison that points to new metric

by Alistair Helm in


I read with interest a joint report by Realestate.co.nz and REINZ (published last week)  “New Zealand Property Report – asking & selling prices - a comparison”. The report states that based on analysis of property sales and property listings in the second half of last year – the median absolute difference between asking price and selling price was 2.67% nationally. That would mean that based on the most recent median sale price of $550,000 the median difference was just $15,000. Clearly indicating a very accurate estimate by agents of likely selling prices.

The report published this chart of asking price to sale price tracking the past 5 years.

I must confess for a couple of minutes I was somewhat confused, as I made the mistake of assuming that what this report had done was to track the monthly asking price as reported by Realestate.co.nz in their monthly NZ Property Report and the monthly REINZ median sale price. The chart for this set of data looks somewhat different as you can see.

The variance of national asking prices vs national sale prices is more like $100,000 as opposed to $15,000. This amounts to a 20% variance as opposed to the reports 2.67%. I then read a bit deeper into the report to understand why I had been confused and thereby explain the significant difference between these two seemingly similar data sets.

This new detailed joint report is based on the relationship between asking price and sales price where a price has been displayed when the property is listed for sale. So the data comprises just those listings where the property has been marketed with a price by the listing agent, thereby excluding all listings by auction, tender, or simply those for which no price is displayed.

Out of interest based on the current portfolio of all listings on the market at this time – the sample set in the report of properties where a price has been displayed when the property is listed for sale is by far the largest subset of properties on the market amounting to 61% of all listings. Some 16,877 from among the 27,643 properties on the market. This data is very helpfully provided on the Realestate.co.nz website under the Advanced Search on the Classic site – unfortunately another weakness of the proposed new website which has no such Advanced Search function.

Screen_Shot_2018-02-13_at_10_47_13_AM.png

Being an analytical person, I began to wonder what this data point of median absolute difference between asking price and selling price was? – was it the amount of the variance of the median asking price to the median sales price for all the listings over that 6 month period? Or was it the median of all the variances between the asking price and selling of all the listings over that 6 month period?

I hope I have not confused you yet!

To hopefully help explain, here are a random set of fictitious data point to help explain my questioning:

comparison_asking_price_to_sales_price_REINZ_Realesatate.png

These 7 properties represent a fairly wide range of prices. The median asking price is $650,000 and the median sale price is $635,000 which relate to property #3. In choosing this fictitious group of 7 properties I have reflected sale prices that are both above and below the advertised price as I assume the listings that feature a price include both those with a price, as well as listings that feature the prefix of “offers over $xx / Buyer interest form $xx / Buyer enquiry over $xx”.

However as you will see the median absolute variance of this data set of 7 properties is not the ($15,000) from property #3 but is ($5,000) from property #4 – with positive and negative variances the median gravitates to a midpoint which in this case is close to zero especially as the extremes of variances are $70,000 below and $55,000 above asking price.

I therefore have to ask – is the use of median absolute variance appropriate?

An alternative data analysis could be to use the mean as opposed to the median. As detailed below the mean asking price to sales price for the same set of properties is $12,000 representing a 1.3% variance as opposed to the 0.9% of the median absolute variance.

comparison_asking_price_to_sales_price_REINZ_Realesatate.png

 

Aside from this question I have with the data point chosen for the analysis, I commend Realestate.co.nz and REINZ for this report. The takeaway is that where properties are marketed with a price; the price chosen at the recommendation of the listing agent is likely to be a very close approximation to the likely value of the property at the time of sale. This is valuable for buyers who often feel they are in the dark regarding prospective value of properties.

As a proposal for these two organisations I would like to recommend an extension of this one-off report. I feel it would be of significant value if Realestate.co.nz started to report this new metric of asking price for new listings that are marketed with a price. Tracking this by region by month as well as backdating data to 2007 would be really valuable extension to the NZ Property Report!


Making sense of online property valuation models

by Alistair Helm in


couple engagement online shutterstock_3467780.jpg

Having cited the democratisation of property data as the most significant event to occur in the real estate industry over the past 3 years, I thought it would be a useful follow up, to provide some insight and perspective as to this new world of more accessible property data and by so doing provide more context as to these estimated valuations as compared to traditional valuation providers.

I do also propose in a follow up article to review each of these new providers and assess their relative strengths and weaknesses.

There are currently I judge five key players in the market offering online estimated valuations for NZ property. These are Homes, Trade Me Property, Realestate.co.nz, MyValocity and QV (a note, QV only provide a free estimated valuation model on their mobile app – their website still requires the purchase of an e-Valuer report at $49.95 per property).

All of these operations provide a free unlimited online automated valuation on pretty much all properties in NZ. Well actually not every property. The fact is all of these providers recognise that without sufficient proximate data from which to compute their algorithm they cannot attribute a reasonable estimate to every property, so not every property will have a valuation estimation. It is likely that the more remote the location, the more rural, the less frequent the number of local sales the less likely there will be for a estimated valuation.

Let me expand upon this as an insight as to how these Automated Valuation Models (AVM) work. Each of these companies leverage the now easily accessible massive computing power that only a few years ago was the reserve of major corporate and government agencies. The likes of Amazon Web Services, Google Cloud Platform and Microsoft Azure to name but a few, which offer massive computing capacity just when you need it – in this case allowing these local companies to rent a couple of hours of grunty computing power to run algorithms that analyse the impact of all recent local sale records for all properties. This is basically how the algorithm works. Each property record is assessed against recent sales of similar properties (similar by standard metrics of for example number of bedrooms, size of property being the two most important).

The key data point here is recent sales; the more recent the sale, then the more accurate the estimation. Naturally there is a lot more sophistication in each company’s algorithm than I have outlined here, including self-learning tools to assess the system’s accuracy by effectively going back and estimating a property sale before the actual sale is confirmed and then reviewing actual sale price against estimation.

Powerpoint_images_2018_WIP.png

AVM’s tend to be displayed on these various platforms as a price range with a mid-point. This is the result of statistical convention rather than a true sense of a predictive range. In my view look at the mid-point of the range as an indication of the AVM rather than the upper price! A range can often be as broad as 15% or even 20% either side of the mid-point which at times makes them seem very inaccurate. The fact is, the computer algorithms compute a single figure together with a confidence factor which then drives the scale of this range.

That is the complex part of Automated Valuation Models (AVM). The key question though is, can you, and should you, trust these estimation valuation models in the marketplace as a guide to better inform you as to an indication as to the likely selling price of a property?

Before I go into that, it is really important to lay out the difference between a number of data points that you are likely to come across in terms of assessing the value of a property. I have detailed below the 5 valuation data points in descending order of accuracy.

The Selling Price – this is ultimate statement of the true value of a property. This is the price at which a willing seller accepted an offer from a willing buyer. This valuation is 100% accurate, but at the same time ephemeral, as it is a moment-in-time judgement and will never be repeated because circumstances with the property market at a hyperlocal level change all the time.

A Registered Valuation – this is the most accurate estimate of a property's value and that is why it is insisted upon by banks and lending institutions who are prepared to take on the risk against which they lend. Registered Valuations are undertaken by a professional valuer, a person who has undergone extensive training and education spanning many years. Such valuations, often cost many hundreds of dollars. Registered valuers use recent sales and local knowledge to provide a very detailed written assessment of what a property is worth in today’s market. There is also professional indemnity that lies behind the valuation report.

A Real Estate Appraisal – a licensed real estate professional will provide a client with an appraisal as to what a property would expect to fetch in today’s market. Under the guiding rules of the Real Estate Agents Act 2008, it is a requirement that a salesperson provide a client with such an appraisal before signing an agreement to list and market their property. Such appraisals need to identify a price or a range, ideally not exceeding 5%. Such an appraisal is computed using a comparative market assessment of what properties of similar size and features have sold for recently. Additionally an appraisal will look at the hyper-local market conditions of supply and demand which a local agent is uniquely able to assess.

An Automated Valuation Model – as outlined above this computer based model is undertaken by the leading five online providers and is based on raw data with no human intervention or local insight.

A Rateable Value – this is a valuation developed for local authorities and undertaken every 3 years in order to provide a benchmark upon which local rates can be assessed. The Rateable Value is judged to be the likely selling price at the time the assessment is made and therefore this estimation decays pretty quickly afterwards. Largely the model used by the providers of this service for the local authorities matches the computer based AVM.

As a prospective buyer or seller the question is, which of these data points should you look at, when and why? Here is my opinion.


What a property seller should do?

Agent and clients shutterstock_165593015.jpg

If you are looking to sell a property it is very useful to keep an eye out for local sales results – many agents nowadays will provide such report at open homes, and of course Homes / Trade Me / Realestate / MyValocity can provide this data although not all provide email alerts of recent sales in your area (Homes does a great job of this). In addition, it does no harm to review the AVM estimate for your own property, it’s a valuable guide. Again Homes offers the ability for you to ‘own’ a property record and receive monthly emails of the latest valuation and market trends.

When you are ready to go to market with your property choose your licensed real estate salesperson and get them to provide an appraisal which will give you their valuation estimation which will be most likely based on selected comparable recent sales of properties that best match your property. Their appraisal report will identify these comparable properties thereby allowing you to discuss and debate the merits of your property versus others. This appraisal is the best indicative valuation you can get without investing in a registered valuation.

The estimated valuation in an appraisal will be either a single figure or a range and in this case best practice says that the range should be no more than 5% overall, which means between 2.5% above and below the mid-point. So for example a range of say from $535,000 to 560,000 would be acceptable.

The appraisal you receive may utilise a couple of well recognised models as well as comparable sales. These being net rate / replacement cost or capitalisation of income. I won’t dwell on these other methods here, aside than to say a professional real estate salesperson will used their skills and knowledge to arrive at an estimated valuation that is the best in the market.


agent buyers shutterstock_171921932.jpg

What a property buyer should do?

If you are looking to buy, I would recommend the same approach of keeping a watchful eye on local sales and see what properties, like the one you fancy buying are selling for. Certainly, review the online service providers to see what the properties that come on the market are valued at based on AVM’s. Trade Me is great at this, in providing a link from most listed properties to the AVM on their Property Insights section.

I would recommend that when you get closer to the decision-making process of buying you chat with the agent for the property you are interested in, and discuss with them the view they hold as to price range and how that may differ from the AVM online – they will be only too keen to share the reasons why they view that their price judgement is more reflective of the local market conditions. Listen closely as they are working every day in the market and their insight is critical.

If as a buyer you require finance on the property you choose, you will likely need a registered valuation as the bank or lending institution will insist upon it, however I would take that lead from the lender rather than rush into requesting a registered valuation before you are certain on the property purchase. Remember obtaining a valuation as part of the financial conditions of a conditional offer for a property is perfectly acceptable.

The one estimated valuation I have omitted to mention in this process is the Rateable Value. The role of the RV has now finally gone. It can finally be ignored and retired from the lexicon of property transactions – interestingly something I suggested back in 2013!

 


Property price trends – a new analysis

by Alistair Helm in


charts analysis and calculator.jpg

I recently examined property sales and listings data in order to measure the clearance rates in the property market as a means to better understand the signals within the market as to trending patterns. In the same vein, I have now turned my attention to property prices as measured in the monthly median sales price by REINZ. This data set has a 25 year history, providing a rich period for analysis.

The graph depicting the past quarter of a century is probably well known and understood by those who follow the market.

Over these 25 years, the median price has with the exception of a few pauses and a single period of decline, edged inexorably upward from the starting level of $105,500 in 1992 (in today’s dollars: $175,500) right up to $550,000 at the end of 2017. That represents a 421% increase over the period, allowing for inflation that is a 213% increase – a more than trebling in median sales prices in 25 years.

Examining this chart can leave one with the misleading impression that that prices over recent years have experienced exponential growth. The reason being that a $50,000 rise in 2017 represents a 10% increase – highly visible on the chart, the same 10% rise in 1992 would amount to just $10,000 barely imperceptible on this axis, creating this impression that recent rises are more significant than a decade or two ago.

I've looked for patterns or trends in the path of median price over this protracted period and judge that the 25 years can be split up into 5 distinct periods as I have outlined on the chart, periods ranging from just over 4 years to 6 years.

I have then separately charted each of these periods. For each distinct period I have deliberately created a Y axis that ranges from a minimum of 20% below the median price at the first month of the period; to a maximum range of 110% above the median price at the first month of the period. This has been undertaken so that each chart can be viewed comparatively with each other.

The interpretation I draw from this analysis of the 5 periods of the NZ property market over the past 25 years based on sale price is that we experience cycles, no great surprise! We've had 3 periods of rises ranging in duration from 50 months to 70 months. Each rise has been followed by a plateau period equally lasting from 62 to 70 months. Within the second plateau period from Nov '07 to Jan '13 was the only significant period of falling prices. This decline lasted 23 months and at the lowest point prices fell 8%.

What is equally striking is the comparison of the three periods of property price inflation - the early 90's and the most recent 59 months both attaining a level of just under 50%, compare that with the staggering 102% rise leading up to the GFC over a period of 70 months. Certainly by this analysis the most recent 5 years have seen strong price inflated but nothing of the extreme seen in the early period of the new century.

For me this analysis proved the value in visualising price movements in terms of relative indexing as I have done with paralleled Y axis in each of the 5 periods. This got me thinking as to how to best represent this indexing in a histogram of property price movements. A bit of experimentation and trial and error has produced this new chart below.

It is a binary chart where the criteria is relative 10 months performance against a base month. It seeks to highlight periods that have experience significant increases in property prices or periods where prices have stagnated or declined - picking out individual months.

By way of demonstration to show how the chart is developed, let me explain. So if as an example the median price in January 2002 is less than the average of the median price in the preceding 10 months then January 2002 is judged to be a month of weak sale price and a red bar is displayed. Similarly taking May 2015 if the median price in that month is greater than 5% above the average median price for the preceding 10 months then May 2015 is judged to a month of strong sales price and a blue bar is displayed. The decision surrounding the use of average rather than max or median; as well as the 10 month period as well as the 5% inflation criteria are purely experimental to deliver what I judge to be a valuable visual representation of the property price trends.

I rather like this representation as a visual cue as to the trend in the market highlighting periods of sustained growth, sustained weakness or variability between growth and weakness.

As to interpretation of this chart and the earlier charts as a guide to the future, I will leave that to you the reader as my role here is not to predict the path of property prices, merely to provide a lens through which to view and make your own judgement as you interpret the data.