Four Reasons to Change the Definition of Recession

The way economists currently define recessions is heavily flawed...

THE OECD made headlines last week when it said Britain's economy was heading back into recession, writes former Bank of England Monetary Policy Committee member Andrew Sentance.

The definition of recession which underpinned their view is widely used in the economics profession – two consecutive quarters of falling GDP. So the fact that the OECD expect a fall in GDP in the first quarter of this year (erroneously in my view) to follow on the fall in the final quarter of last year means that we might "technically" be back in recession. (See my latest PwC Economics in Business blog entry for more discussion on this issue.)

I have never liked this technical definition of recession. It seemed to become more widely used in the UK in the 1980s and 1990s, and its origins can be traced back to the United States in the 1970s.

There are four problems with it. First, the time period is relatively short. Two quarters is a period of six months, but data on GDP in one quarter can be heavily influenced by what happens in a single month. In December 2010, a bad bout of snow brought the UK economy grinding to a halt and knocked 0.5% off UK GDP according to the Office for National Statistics. And there is now an active discussion about whether the burst of panic buying of fuel, jerry cans and other items could materially influence the Q1 GDP figure for this year. (See, for example, Andrew Lilico's Telegraph blog)

Second, GDP figures are frequently revised. So the snap assessment which is made on the basis of one set of figures can be reversed by later revisions. This is particularly so for the UK. In the early 1990s, the recession appeared to continue through 1992 and into early 1993. At the time, it was believed to be the longest recession that the UK had experienced since the Second World War. The most recent data show the decline in GDP coming to an end in late 1991, exactly when Norman Lamont declared he could see the "green shoots of recovery". He was right, though ridiculed at the time. At the IEA State of the Economy conference in February, Norman Lamont himself noted that the GDP figures covering his time as Chancellor were now unrecognizable from the data available at the time.

Third, I believe it is wrong to base your judgment of the state of the economy on a single economic indicator. GDP probably contains some useful information about the current state of the economy. But so do the employment numbers, business surveys, retail sales and a raft of other data. In the autumn of 2009, this broader picture of the economy suggested to me that the UK economy was recovering (see the speech I gave at Royal Holloway in November 2009), even though the ONS estimates of GDP at that time showed little sign of sustained recovery. According to the initial estimates, GDP fell in the third quarter of 2009 and rose by just 0.1% in Q4. The revised figures are now much more consistent with the view I formed looking at a wider range of data, with growth resuming in Q3 and strengthening in Q4.

The final problem is that a slow-growing economy runs a greater risk of falling into recession than a fast-growing one, just because its underlying growth rate is closer to zero and risks falling below it. That does not necessarily make sense because economies vary greatly in terms of their underlying growth rates. China – which grows at 6-10% per annum – would have to suffer a cataclysmic economic accident to experience a recession on the technical definition. Yet economies with declining populations and low productivity growth – like Japan – will appear excessively recession-prone.

So what can we do to come up with a better approach to defining recessions? One suggestion is to form a committee to do the job. In the United States, the National Bureau of Economic Research (NBER) has the technical role of defining turning points in the economic cycle. However, they come out with their judgments after quite a delay. Not particularly helpful for business and general public awareness of the state of the economy. And the notion of setting up another Committee to judge on economic progress (we already have the MPC, OBR and the Treasury Select Committee) does not appeal.

A combination of simple indicators seem to be the answer. The three indicators I would choose are: (1) an underlying measure of GDP, such as non-oil GDP or a weighted average of manufacturing and services output; (2) the unemployment rate; (3) a weighted average of reliable business surveys of business activity. However, in terms of GDP, the "two quarters of decline" rule does not make sense. There should be some reference to the underlying trend rate of growth and how far GDP has dropped below it.

If you have three indicators, it may not be necessary that all three are signaling recession – particularly recognizing the unreliability of early GDP data. Perhaps we need to take some advice from Meat Loaf, who proclaimed in 1977 that "Two out of three ain't bad". Or in this case two out of three is bad – recognizing that a recession might be a "Bat out of Hell"! If two out of the three indicators are signaling recession – then we are likely to be in one. But that is a long way from the so-called "technical definition" based on dodgy initial GDP estimates which so many people now casually and unthinkingly use

Investing in Gold? See how BullionVault can cut your costs of buying the safest gold on the planet...