This time at the AFR following the MB lead:
Chris Curtis, a Sydney buyer’s agent who has acquired hundreds of millions of dollars of property on behalf of clients over the past 10 years, is scathing about the quality of information in the market more broadly.
He says the primary problem is that much of the raw information – including heavily reported weekend auction results – is provided to research companies by people within the real estate industry, primarily agents who have “a vested interest in one side of that transaction and in conveying the impression of a consistently buoyant market”.
…”It’s more slanted to higher-end properties,” says CoreLogic head of research Tim Lawless. “It reflects what’s happening in premium markets.”
…Done properly, hedonic models are the gold standard of house-price measurement. However, they are underpinned by a number of assumptions – such as values ascribed to certain attributes – that are not always clear. A tweak in the underlying algorithms – akin to changes Google makes to its search engine – can significantly alter the results it comes up with and the picture may not be clear to outsiders. And then there is the issue of how open a data provider is.
This drawback was clear last month when the RBA called out CoreLogic for having “overstated” housing price growth. It turned out the housing provider had indeed tweaked its model without telling some key parties – including the RBA – to adjust the way it accounted for extremely high- and low-priced property sales. The change meant it exaggerated the vigour of strong markets like Sydney and Melbourne – and inflated the aggregate increase – and also overstated the weakness of a soft market like Perth.
CoreLogic said it would restate its index to account for the change, but it hasn’t yet done so. It declined to supply the AFR Weekend with an estimate of the impact of the restatement. The change affected its figures in April and May, so it will only be from June 2017 – by when April and May 2016 have dropped out of the equation – that CoreLogic’s annual price growth figures won’t be affected by the distortion.
CoreLogic is doing an audit to ensure its methodological changes are sound before re-releasing the index, which is sensible enough even if it raises the obvious question of why wasn’t such done before the change was made at all?
To err is human so that’s forgivable, but let’s have it!
