Thursday, 6 March 2014

The ONS new treatment of Investment (wonkish)

Hats off to the ONS, who had a special session to the public on Wednesday on their new investment data (which the Bank of England have publicly cast aspersions on).  Here's my take.

First, off real investment is revised down, GFCF on the left, whcih includes dwellings and business invsetment on the right, which has not fallen so much.

Underlying this is a big chance in the distribution of nominal investment in the underlying assets, with a big rise in intangibles.

what's causing this?
  1. software has been taken out of machinery and equipment.   Software is bundled with hardware and so this is hard, but if it's done right, then machinery/equip is now just hardware plus non-hardware machinery/equip.
  2. there's a new method for calcualating artistic orginals, done by Peter Goodridge at Imperial.
Putting this together, gives a remarkable nominal investment picture:

Ignoring the jump in 1995 for intangibles, which is a data issue, we see that intangibles are now close to plant investment and closing on buildings.  Note that when R&D is captialised, in the summer, that will add about £15bn to intangibles which will push intangibles up to close to the highest item.

But what about real? Here there is a problem.  When software is taken out of plant its nominal share falls.  An overall investmetn price index is a nominal share weighted index of the underlying asset prices.  When software was in plant, it was then as if software prices were falling since plant prices are falling a lot.  But when software is taken out, then ONS have to give it a deflator. They have chosen one that is rather flat, not falling: the US have, more or less, an index that falls.  So if you think the software price is falling, then the old price index was right and the new one wrong, despite the improvement in the method.