The August housing starts number was a disappointment, printing at 750K on expectations of a rise to 767K from last month's 746K, now revised lower to 733K. This would have been a boost to a market trained to expect more QE on any economic weakness, if only all QE in perpetuity, and certainly at leat $85 billion in monthly flow, was not already priced in. As a result, we are slowly getting to the dreaded point where bad news is once again bad news, at which all faith in the Fed as a monetary policy vehicle is lost (since Fiscal policy is now perpetually deadlocked). If there was any good news, it was in the single family starts which printed at 535K in August, a rise of 28K from July, and the highest since April 2010 (when housing had again "bottomed") driven by a surge in new building in the Midwest to 134K, from 111K. Finally housing permits which are nothing but noise, declined but beat expectations modestly. Since permits are a completely meaningless category and are purely used by hedge funds to game the market (they cost a token amount of money to procure, involve no actual work, and are there merely to frame the "housing has bottomed" trope time after time, until disproven), just like Libor, there is no point to observe them.
Source: Census Bureau