Join our community of smart investors
OPINION

GDP and statistical fetishism

GDP and statistical fetishism
January 18, 2012
GDP and statistical fetishism

Peter Spencer of Ernst & Young’s Item Club says the economy is “probably in technical recession at the moment.” Economists at Fathom Consulting agree. They think GDP fell by 0.3 per cent in the final quarter of last year. But Simon Kirby at the NIESR disagrees. He estimates that GDP grew by 0.1 per cent in Q4. “The recession is over.” He says.

You might think this dispute will be settled next Wednesday, when the Office for National Statistics releases its first estimate of GDP in Q4.

You’d be wrong. Those numbers won’t tell us what GDP did. They’ll tell us what statisticians with partial and imperfect information believe it did. There’s a difference. On average, the first estimate of quarterly GDP growth has under-reported growth by 0.1 percentage points since 1997. This is in itself inconsequential, but it is the average of over- and under-estimates. If we ignore the sign, the average error in ONS’s first estimate of GDP growth has been 0.4 percentage points. And this is just an average. Some errors have been larger. For example, the first estimate for 2008Q2 was that the economy grew by 0.2 per cent. But the ONS now believes it contracted by one per cent.

Such errors – or more precisely, proneness to revision – mean that next week’s number won’t tell us much, unless it’s a huge surprise. A report of a small fall will be consistent with GDP actually growing. And a report of a small rise will be consistent with it contracting.

This is not the only reason why I don’t care about the GDP numbers. Whether the numbers – even accurate ones - show a small rise or small fall in GDP, the big picture is the same. The economy is doing poorly – it is “broadly flat” as Bank of England Governor Sir Mervyn King said with all the precision we need.

What really matters is not a statistical construct but the experience of real people. And we know from today’s labour market report that this is grisly; full-time employment is falling; unemployment is rising*; real wages are falling for those in work; and there are 6.3 million people who would like to work more than they are.

The economy is not doing well enough to improve living standards. That’s what matters.

Worrying about whether it is up or down by a fraction of a percentage point is mere statistical fetishism. When the economy was growing steadily before 2008, nobody much cared whether the quarterly GDP numbers showed a rise of 0.5 or 0.7 per cent. So why obsess over plus 0.1 or minus 0.1 per cent? Nothing of substance hangs on it. (In saying this, I do not disparage the NIESR or Fathom Consulting; they are agreed on the fact that our economy is in poor shape.)

But here’s a forecast. If the ONS does report a small fall in GDP, it’ll probably be the lead story for much of the media: “Britain back in recession.” However, such stories will tell us less about the economy than about journalists’ attitudes to numbers - which is like their attitude to sausages: they love them but don’t much care how they’re made.

* Actually, we can’t even say this with certainty, as there’s a wide margin of sampling error in employment data.