From Zero Hedge
Last month, when, with great amusement, we reported that "New Home Sales Explode Higher Thanks To... Record High Average New Home Prices?", we mocked the latest batch of bullshit data released by the US department of truth as follows:
New Home Sales rose a magnificent (seasonally-adjusted annualized rate) 18% in August - the biggest monthly rise since January 1992 albeit with a 16.3 90% confidence interval, meaning the final number may well be +1.7%. At 504k, new home sales are back at May 2008 levels (though obviously massively below the 1.4 million homes sold at the peak in 2005). As a reminder, May's 504K new home sales print was later revised later to 458K. But even more stunning, new home sales in The West rose a mind-numbing 50% in August (and up 84.4% YoY - nearly double).
Well, it is now a month later, and here come the revisions: first, that 50% surge in the West was revised... 30K lower. But to get a sense of just how bad the revision was, here is the old, pre-revision data, and the "data" following the latest revision.
In short: the euphoric, consensus-beating data for every single month since May has been revised lower, by on average 6% and as much as 9%. Perhaps finally people will realize that there is only one number that matters in the Census bureau's monthly new home sales report: the ±15.7 90% confidence interval. Well, people maybe, but not algos, who only care about one thing: whether the data beat or missed.
Now we wonder: will all those market surges over the past 4 months which were based on erroneous headline data, all be revised lower? Sarcasm off.
Oh, and as for that record new home price reported last month, which magically also resulted in what the US government wanted everyone to believe was a surge in buying... well, see for yourselves:
So to summarize: the latest "housing recovery" has been indefinitely postponed due to data revisions.