by Calculated Risk on 7/27/2006 04:21:00 PM

UPDATE: I've checked more data, and historically revisions can go either way - although negative revisions are more common, especially recently.

As Ed Hopper and Robert Coté mentioned in the comments of the previous post, the New Home Sales number has been consistently revised downwards. I usually note this every month.


Click on graph for larger image.

This graph shows the initial release estimate for New Home Sales (red) followed by the three revisions.

This pattern can also be seen in data for 2004 and 2005 (I haven't checked all data). For whatever reason, the initial release typically overstates New Home Sales - and it is reasonable to guess that the final number (after three revisions) will be around 5% less than the initially reported number.

The final number for June will probably be close to 1.075 Million, a decrease of 14.5% from June 2005, and the lowest June since 2002.