Posted by Consistent with their daily series, the RPData house price series for June showed a decent bounce of ~1%m/m. The headline NSA numbers were +1%m/m for the 8 Capital City number, with houses +1.2%m/m and units -0.3%m/m.
Some have reported that the seasonally adjusted increase was 1.4%m/m – saying that house prices rose against seasonal headwinds – but I am not so sure.
I did the seasonal adjustment in my usual way, allowing the seasonal factors to vary over time, and I got a weaker seasonally adjusted increase (+0.4%m/m) — which was something of surprise to me as those saying 1.4%m/m SA’d are credible. Also, I don’t look closely every month, but I knew that the seasonal factors had used to be reliably negative in June.
The above chart shows the seasonal factor splits (all log-additive). The dark black line uses a ‘stiff’ method, which makes the seasonal factors all equal across the entire sample (1995-12 to 2012-6) for each month; the green line is a more flexible estimate of the most recent monthly seasonal factors (year to June 2012), and the blue line is the same from five years prior (year to June 2007).
It looks to me that the seasonal dynamics have changed. June is no longer a weak month (note this is for the aggregate 8 Capital City Dwelling price series – if you SA component by component and aggregate you will almost certainly get a different result).
Using my SA, I see nationwide unit prices down a titch (see above), but house prices up 0.5%m/m (see below).
The aggregate SA’d index (note, i seasonally adjust the aggregate number, and do not sum the house and unit indices) rose by ~0.4%m/m.
The Sydney market appears to have got some pep from the rate cuts — with unit prices rising a SA’d 0.8%m/m (it’s in the unit market that recent cuts seem to be working) …
and house prices +0.6%m/m following a terrible May.
The Melbourne market is less impressive, with unit prices falling an SA’d 1.1%m/m (to make a ~5% drop in the prior 3m)
… happily, however, the Melbourne housing market appears to have firmed, with prices rising +0.3%m/m, but that’s following an ~5% drop over the prior six months.
Finally a comment about the data. When I see data like this — the swings up and down of ~1% per month for the last two months when the underlying price process is probably smoother than that — it is typically either a sign that we are at a turning point, or an indiction that there has been some measurement error.
Folks are making both claims right now — I don’t think that there’s enough evidence to pound the table on either side. We’ll just have to wait and see if house prices have bottomed.
ricardian ambivalence 