Prediction Precision

‘Prediction is very difficult, especially if it’s about the future.’ – Niels Bohr.

Although that crystal ball can be a bit foggy at times, it doesn’t mean we can’t make educated judgements on the likelihood of future outcomes. It does, however, highlight the folly of making predictions with the certitude found in many property market reports today.

The BIS Oxford Economics’ Residential Property Prospects 2017 to 2020 is one such report. The debater in me would like nothing more than to parse each individual line in the report, but that’ll take too long and wouldn’t be productive. Here are the key predictions for 2017 to 2020:

  • Sydney: house and apartment prices are expected to fall 4% by 2020

The facts:

It costs 43% of a monthly household income to service a loan on an average property in Sydney. If interest rates rise just 1%, it would cost 55%. Couple that with a record high price-to-income ratio and it is likely that house and apartment prices will remain stagnant in the medium term, as they did between 2004 and 2009.

  • Brisbane: house prices will increase 7%, while unit prices will fall 7%

The facts:

We expect Queensland to benefit significantly from the surge in gas exports. Looking further ahead, rising export earnings will help tilt the growth pendulum back to Queensland from 2018 onward.

In the 2015-16 financial year Brisbane recorded the highest internal migration of any city, with a gain of 10,149 people. Brisbane’s largest intake of interstate migrants came from Sydney, with a total of 9,900. Sydney lost a total of 23,176 over the same period. What about that pesky supply? Rising construction costs, increasingly stringent lending and a reduction in foreign investment will see the number of dwelling completions evaporate almost entirely by 2019.

The city’s economic fundamentals, particularly the near doubling of gas prices and strong population growth, will see investment increase and unemployment decrease, both of which are positive for property markets.

  • Melbourne: growth of 5% for houses, apartments predicted to fall 4%

The facts:

Segmented market: property prices in much of the north and west of Melbourne remain well below the capital city average. Given nation-leading figures on population growth, economic growth and job creation, our research indicates continued demand for property, particularly in the more affordable segment of the market. Read more here.

Prediction vs Reality

In 2015, BIS predicted that ‘over the next three years, Melbourne’s median apartment price is forecast to fall by 12%’. The reality is that between March 2015 and December 2016, the median apartment price increased by 8.4%: that’s more than a 20% turn-around. House prices were forecast to increase by 5% in the two years to December 2016. The reality? 30%.

What did they have to say for Sydney?

‘Sydney prices have run well ahead of Melbourne in the past two to three years and Sydney prices being more strained from an affordability perspective, probably have a bit more downside once things start turning.’

The reality is that Sydney house prices increased by 25% in the two years to December 2016.

Prediction is a very difficult game indeed, but that doesn’t mean you go in blindly and play the property edition of ‘pin the tail on the donkey’ to select an asset. An asset class worth more than six trillion dollars deserves more respect than that. Rather than make bold forecasts down to the per cent, we assess a property market on the extent to which it exhibits the economic and demographic fundamentals conducive to demand. Those that claim powers to do more are either psychic or overstating their abilities.


1st Oct
Most people die at 25. We just bury them at 75
24th Sep
Israel and Hezbollah conflict and RBA rates decision
17th Sep
Bite Sized Basics: Thinking Beyond the Obvious – Henry Ford’s Insightful Quote
There are no results to display. Please try a different keyword or reset the filters to see everything.

Subscribe for free property investment advice, resources & education

This field is for validation purposes and should be left unchanged.