One way of trying to solve the mystery of the extreme weakness in construction this year - the main drag apart from special factors on the GDP numbers - is to dig deeper into the statistics.
Paul Bivand of the Centre for Economic and Social Inclusion (usually known as Inclusion) has done this by looking at the detailed employment numbers for the construction sector. They show that while employment demand weakened significantly at the start of the year, supporting the weakness of production at that time, it has since strengthened significantly.
There may be, as Bivand says, a rational explanation, but just as likely is that the construction figures used in the GDP calculations are implausibly weak. His note is here. One thing we do know is that the construction revisions are extraordinary. An initial estimate of a 4.7% drop in the sector's output in the first quarter of 2011 has now become a 0.5% rise.