Britain never had a double dip recession – and that doesn't matter

The magic of GDP revisions.

The ONS has released a bumper set of GDP revisions, taking a fresh look at the national output all the way back to 1997. The headline points:

  • The recession in 2009 was much more severe than previously thought. The agency has revised its estimate of the 2009 contraction from 4 per cent of GDP to 5.2 per cent; an increase of a third.
  • A change of 0.1 percentage points has erased the "double-dip" recession in 2011. The contractions in Q4 2011 and Q2 2012 still exist, but Q1 2012 is now estimated to have had exactly 0.0 per cent growth; the wiggle thus no longer qualifies as a technical recession.
  • Historically, the 2001 slump after the dot-com boom also appears to have been much worse than previously thought. Rather than the economy growing by 2.9 per cent that year, it grew by 2.2 per cent, a revision of 0.7 percentage points.

The immediate effect of the revision is political. The double-dip recession has been extraordinarily damaging for the Government, allowing the Opposition a concrete fact to point to to back up the claims that they "inherited growth and provided recession". The revisions eliminate the negative growth in one quarter, and mean that that period no longer meets the requirement of two or more consecutive quarters of negative growth which defines a recession.

But that is a technicality of the highest order. The estimate for growth over those three quarters combined is exactly the same as it was yesterday: a contraction of 0.6 per cent. All that has happened is that the timing of some of that contraction has shifted. Q2 2012 is now thought to be marginally worse, and Q1 2012 is now thought to be marginally better.

In other words, as I said back in April 2012, technical recessions don't matter. They impose an artificial dividing line between "good" and "bad" growth, when in fact, going by yesterday's stats or today's, that period is and always was mediocre. Horribly, cripplingly mediocre.

Anaemic growth is just as bad as a technical recession. In many ways, it's worse; whereas technical recessions are at least expected to end, the stagnation which we are living at the moment has stretched out for almost half a decade, and could well last almost half a decade more.

The more interesting fact is that the recession and depression have been massively increased in size. Since 2008, the economy has actually contracted by 1.1 per cent more than we thought. That increases the so-called "productivity puzzle": why has unemployment been falling while GDP hasn't been increasing? That puzzle is partially solved with the realisation that the labour market improvement stopped six months ago; but there still appears to be a disconnect between the two measures. The commonly cited solution to the puzzle is to argue that labour productivity has dropped; but even that just pushes the question further down the tree. Why is our productivity down so much?

But the number one take-home lesson from all of this is: GDP is actually kind of crappy. It's estimated from samples several steps removed from actually "measuring the size of the economy". It misses out huge sectors of the economy, from unpaid caring and parenting to increasingly mainstream aspects of the shadow economy like transactions in bitcoin. And it's wrong; sometimes for decades on end.

Perhaps we should take the advice of former MPC member Andrew Sentence (who has been predicting this recession would be revised out of existence for a year), and focus on more important and less volatile indicators like unemployment and business activity. Not only would that reduce the risk of rewriting history: it would also mean that doing well could no longer be defined as "not failing".