Very weak jobs data from Statistics New Zealand yesterday raised economists' eyebrows.
They see the labour market as soft, but not as weak as a 1 per cent drop in employment in just three months would imply.
The unemployment rate dropped to 6.9 per cent from 7.3 per cent in September because of a collapse in labour force participation - the proportion of the working age population either employed or actively seeking work. It recorded its steepest fall and to the lowest levels since 2004.
Had the participation rate remained unchanged, the unemployment rate would have been an eye-watering 8.4 per cent, the highest since 1994, ANZ economist Sharon Zollner said.
"The fact that the number of people receiving an unemployment benefit is yet to rise strongly suggests that number does not reflect reality."
So why does this survey of 15,000 households tell us the number of people employed fell 1.4 per cent last year, when a survey of 18,000 workplaces recorded a 1.4 per cent rise in filled jobs?
Is it plausible that 47,000 fewer people are employed outside Canterbury than a year ago, a drop of 2.5 per cent, when the economy is thought to have grown around 2 per cent over that period?
Or that the self-employed dropped by 50,000 or nearly a fifth, over the past year to an 11-year low?
And if employment is in decline, how come the PAYE tax take is up nearly 7 per cent on a year ago?
The Bank of New Zealand's head of research, Stephen Toplis, points to other puzzles such as the contrast between the contraction in employment recorded by the household labour force survey (HLFS) and the hiring intentions in surveys of business sentiment.
"Why have businesses been telling us porkies for so long and why do they continue to expect employment growth?" he said.
"Why are consumers so confident when so many are supposedly losing their jobs? The housing market usually reflects the state of the labour market. Why is it so strong?"
The HLFS is a survey and subject to sampling error. Strictly speaking Statistics NZ is saying it is 95 per cent confident employment in the December quarter fell 22,000 to 2,194,000 - give or take 22,100. The rise of 10,000 unemployed is also within the margin of error. About 40 per cent of the drop in the labour force occurred among those over 65. There was also, Zollner notes, a sharp increase in 15- to 19-year-olds in education.
"This may be a bad sign with respect to the current state of the labour market but it is positive for New Zealand's long-term growth prospects," she said.
The survey has construction workers down 4.3 per cent on a year ago, which is hard to square with the rebuilding in Canterbury gathering pace.
Zollner said the HLFS surveyed established households and would not pick up people in temporary accommodation, probably involved in in the rebuilding.
As for the disparity with Tuesday's quarterly employment survey (QES), based on data collected from businesses which recorded 0.4 per cent growth in employment in the December quarter, Statistics NZ's labour market survey project manager Daniel Griffiths points to several respects in which the QES is less comprehensive than the household labour force survey.
It does not include agriculture, the self-employed or those working unpaid in family business. It also misses the 90,000 people working in enterprises with just one or two employees.
If those groups are stripped out of the HLFS, employment rose over 2012, not too far from the 1.4 per cent rise in filled jobs recorded by the QES.
The latest quarter's drop in employment was entirely explained by a drop of 31,000, or 6 per cent, in people working part time (less than 30 hours a week). Full-time employment rose 0.4 per cent and usual hours worked rose 0.3 per cent.
Over the year the number of people working between one and 15 hours fell by 20,000, but the number who usually work over 40 hours a week rose by 30,000, Griffiths said.
Given the progressivity of the income tax scale, that would help explain the apparent disparity with the PAYE take.