Fact check: Business confidence surveys have little to do with actual economy

Wednesday, 11 April 2018

Sometimes the economy can be trucking along fine while businesses supposedly panic.

ANALYSIS: A week or so ago ANZ's research division released their monthly business confidence numbers, and as ever they made news in basically every place they could.

A net 20 per cent of the businesses surveyed felt negative or pessimistic about the year ahead, leading to an eye-opening '-20' rating. This was down on the month before, but not as far down as it had been right after the Government took office, when it plummeted to -37.8.

Journalists love a series of comparable numbers, but politicians love them even more. National's Finance Spokeswoman Amy Adams told the House the numbers were falling because the Government was 'shooting from the hip' and had ' zero ideas, zero competence, and seems to have no plan of where it's going'.

It's unsurprising these numbers become political football. But even ANZ admit that the headline business confidence figures don't actually correlate very strongly to GDP, a much firmer measure of how the economy is actually doing. In their March release they noted that 'Firms' views of their own activity (which has the stronger correlation with GDP growth), lifted from +20 to +22.'

* Slump in business confidence poses 'material risk' to economy

* Economists see government debt rising billions more than Labour's plan**

Former Reserve Bank economist Rodney Dickens of Strategic Risks Analysis believes the survey has a major political bias. Basically business leaders are likely National Party supporters and this view biases them against the new Government more than any actual concrete business risk.

Luckily, ANZ has been running this survey for decades so we can compare actual confidence with GDP quite well. Dickens' analysis (above) shows that since 2002 the business confidence survey had just 0.2 correlation with actual annualised GDP growth in the quarter. Basically the businesses got it right just 20 per cent of the time.

'It provided really negative readings for periods that would usually be consistent with the economy being in recession when the economy was actually growing robustly,' Dickens said.

This big change happened after the 2002 election brought in some major changes from the Labour government the business community didn't like, Dickens said.

'The survey became totally useless as an indicator of economics and was swamped by political bias.'

However, if you look at another part of the survey which asks about how businesses are feeling about their own businesses - not the wider economy - the numbers are much more accurate, with a 0.53 correlation since 2002.

And as for the claims of political bias? Well here's the headline business confidence figures with each Government split off.

This is obviously a point the Government has made plenty, asking patsy questions in the house about the correlation between GDP and business confidence in an attempt to take the wind out of National's sails. And to be fair to ANZ, they can hardly control for the political bias of the business community - they are the people best-placed to know about how business is doing. But next time you see a headline about business confidence it might pay to whip out a few hefty grains of salt.