Rodney Dickens says the business confidence surveys have major political bias, and he checks how accurately these surveys signal growth trends

By Rodney Dickens*

The ANZ business survey has gone AWOL again.

Unfortunately, the bank economists and therefore the media haven’t focused enough on the massive political bias that currently overwhelms the results of the survey and make it a poor input into business and investment decisions.

The NZIER business survey has been corrupted less by political gamesmanship but some components have become poor leading indicators while even the most useful component is under a cloud.

This Raving puts the two main components of the ANZ and NZIER surveys in context which is critical for assessing whether much if any weight should be put on them in making business and investment decisions.

When what were once useful leading indicators can’t be relied on it increases the importance of having access to quality analysis of economic and housing prospects as input into business and investment decisions.

The ANZ business survey has gone AWOL again

Based on the ANZ business confidence survey the economy is heading for a recession (i.e. negative GDP growth) with significantly more firms negative than positive (left chart). Based on the ANZ own activity survey near-term prospects for annual GDP growth aren’t so bad but growth should slow to below 2% over the first half of this year (right chart).

When these surveys tumbled in November after the election outcome was finalised the results were reported as if there was a significant threat to economic growth.

For example: “Business confidence tumbles to 8-year low in November”  and “Slump in business confidence poses ‘material risk’ to economy

You can’t entirely blame the journalists for reporting the survey results verbatim. They lack the resource and maybe knowhow to back-test the reliability of the surveys. Consequently, they rely on the bank economists to provide the appropriate interpretation, which is a mistake.

The headlines should have read:
“Businesses again hijack survey to make political protest”
“Survey greatly overstates economic risks from the change in government”

Prior to 2002 these surveys were useful indicators of near-term economic growth prospects as can be seen in the two charts above. Prior to 2002 the ANZ business confidence survey had a 0.7 correlation with annual GDP growth based on it being advanced or leading by three quarters while the ANZ own activity survey had a 0.8 correlation with GDP growth based on it being advanced or leading by two quarters. A 0.7 correlation out of a maximum possible 1.0 is akin to a 70% mark in an exam while a 0.8 correlation can be compared to an 80% mark (i.e. they were quite useful leading indicators of GDP growth).

However, since 2002 the ANZ business confidence survey has only 0.2 correlation with GDP growth based on leading by three quarters (i.e. akin to a 20% mark) while the own activity survey has a somewhat more respectable 0.53 correlation (i.e. like a 53% mark which is barely a pass). The coloured arrows in the two charts approximately reflect the periods when National and Labour were in power. It is more the case with the business confidence survey, but both significantly understated near-term GDP growth prospects for much of the period when Labour was in government and overstated prospects for much of the period when National was in government. When in early 2016 Reserve Bank Governor Wheeler effectively asked for feedback about the negative impact of the fall in dairy farm incomes on the economy both surveys conveniently tumbled temporarily; providing him the evidence he was seeking to justify OCR cuts.

In my assessment the ANZ economists should have pointed the political track record of these surveys when they reported the November and December results but instead they chose to downplay the relevance of the change in government on the surveys. Or, more correctly, the likelihood that the surveys had gone AWOL again meaning they should be largely ignored as leading indicators of GDP growth.

The NZIER survey hasn’t been corrupted as much by political gamesmanship

Thankfully the commentary by the NZIER principal economist regarding the falls in many components of the NZIER quarterly business opinion survey was more balanced. This has in turn largely been reflected in the media commentary on the results of the December quarter survey (two quotes below and related links).
Business confidence drops, with fewer expecting to hire or invest” and “NZIER business opinion survey shows the usual fall in confidence after a Labour-led government takes office, effect of election on actual business activity muted.

But there is still a lack of supporting analysis to put the NZIER December quarter results in context and in my assessment there hasn’t been enough focus on political gamesmanship; instead, like the commentary accompanying the ANZ November and December surveys, the initial focus was on “uncertainty over new Government policies”.

The left chart shows the NZIER business confidence survey has, like the ANZ business confidence survey, become much less useful as a leading indicator of economic growth since 2002. As the NZIER principal economist goes on to point out, it has generally had a negative bias while Labour governed and a positive bias while National government although this is only been the case after 2002.

By contrast, the NZIER own activity survey has a slightly higher correlation with GDP growth after 2002 than prior to 2002 (right chart below). The NZIER own activity survey has a peak correlation based on it leading annual GDP growth by one quarter although is this effectively means a two quarter lead given that GDP numbers take a quarter to be released.

Consequently, the fall in the NZIER own activity survey in the last two quarters should possibly be taken reasonably seriously in terms of what it indicates for near-term GDP growth. However, one qualification is that over the last year this survey has significantly overstated near-term GDP growth prospects. If that remains a case it means GDP growth will slow more in the near-term. A second qualification is that the latest fall in this survey may reflect at least an element of political bias.

When you can’t rely on the insights provided by some of the business surveys and some question marks hang over even the most useful of them it increases the importance of making business and investment decisions based on quality analysis of economic and housing prospects. Our driver-based approach to forecasting that is supported by enlightened use of leading indicators rather than blind faith in indicators that have gone AWOL should be a must have for any businesses and investors wanting to make informed and profitable decisions.

*Rodney Dickens is the managing director and chief research officer of Strategic Risk Analysis Limited.