Tax-based data on employment and earnings paint a less positive picture of the labour market than the official statistics released last month.
They provide a reality check that should reinforce the Reserve Bank's messaging that it is going to be a while before it contemplates easing back on monetary stimulus.
The official labour market statistics released on February 3 recorded a drop in the unemployment rate to 4.9 per cent in the December 2020 quarter, down from 5.3 per cent in September. It was startlingly better than the 5.6 per cent consensus expectation among forecasters, including the central bank.
Nor was it flattered by any surge in people withdrawing from the labour market. The participation rate, which is the share of the population of working age either employed or actively seeking work, held up above 70 per cent.
Instead, the unexpectedly good unemployment rate, we were told, reflected an increase in the number of people employed, up 0.6 per cent or 17,000, seasonally adjusted, from the September quarter and up 0.7 per cent (19,000 people) from December 2019.
Again, this was at odds with market expectations of little or no increase in employment. Those expectations were underpinned by the conventional wisdom that the labour market is a cyclical laggard, slow to respond to downturns and upturns in the economic cycle, and by the fact that wage subsidies largely expired during the quarter.
The official employment numbers are derived from Statistics NZ's household labour force survey (HLFS). It is a large survey, of around 15,000 households, an eighth of which turns over in any quarter.
Care is taken to keep it as representative of the whole population as possible. But inevitably it is subject to sampling error, especially when sliced and diced.
By contrast, the PAYE-based data Statistics NZ released this week is based on the information firms are required to file to Inland Revenue every payday. That is every firm, not a sample of them.
And they show headcount employment ("filled jobs") in the December 2020 quarter was down by 0.9 per cent on the same period in 2019. Not up 0.7 per cent.
That's down by 19,000 people, not up 19,000 people.
A couple of things could account for the difference from the official labour market statistics.
The HLFS count of the employed includes employers, the self-employed with no employees, and unpaid family workers. The tax-based figures do not. That helps explain why it puts the number of employed at 2.7 million while the PAYE-based data has 2.1 million.
The HLFS recorded a remarkable increase of 30,000, or nearly 10 per cent, in the number of self-employed people with no employees compared with a year earlier.
On the other hand it does not include, as the PAYE-based data do, filled jobs held by people who do not live in New Zealand long term, such as seasonal horticultural workers or those here on working holiday visas.
In normal times those temporary workers represent a significant share of the workforce.
In a Cabinet paper presented in December 2018, officials reckoned that at any given time about 170,000 people were living and working in New Zealand on a temporary work visa.
They represented more than 6 per cent of the total labour force but, as the Cabinet paper noted, they are particularly to be found in a few lower-wage sectors, where their share of employment is closer to 20 per cent: accommodation and food services; agriculture, forestry and fishing; and one called "administrative and support services", which includes such things as cleaning, pest control, gardening and packaging products.
The closure of the border over the past year has clearly disrupted the normal flows of temporary foreign workers.
It helps explain why the PAYE data has the number of people employed in accommodation and food services down 7 per cent, year on year, and in administrative and support services down 6.9 per cent. Employment in agriculture, forestry and fishing, on the other hand, was up 1.8 per cent.
The biggest gains by industry were in construction, up 3 per cent, and public administration and safety, up 9.8 per cent.
As for gender disparities, the conventional story that women have taken the brunt of the downturn is not borne out. Compared with December 2019, male employment is down 10,000 and female employment down 9000, in both cases a decline of 0.9 per cent.
It is true that in the June and September quarters female employment on an annual basis fell more sharply than male employment, but that followed six quarters when female employment growth outpaced male by significant margins.
The regional breakdown shows annual declines in employment levels for all of the South Island apart from the not very populous Tasman and West Coast, while the North Island saw employment rise with the important exception of Auckland, where it fell 2.1 per cent.
For businesses chasing the consumer's, dollar it is aggregate wage and salary earnings that matter, rather than headcount measures of employment.
Here again the tax-based data tell a more sobering story than the official numbers based on the quarterly employment survey (QES) of firms.
The QES recorded a 4.5 per cent annual increase in total weekly gross earnings, of which 1.2 percentage points was the increase in full-time equivalent (FTE) employees and 3.3 percentage points the increase in average weekly earnings per FTE worker. The Reserve Bank sees the latter as reflecting a shift in the composition of employment from lower- to better-paid jobs, rather than a general increase in real wages.
By contrast, the PAYE data recorded a 1.2 per cent decline in total earnings compared with a year earlier, a stark turnaround from increases of more than 7 per cent in the two prior years.
It included double-digit falls in aggregate earnings of workers in the transport, postal and warehousing industry and in arts and recreation.
These numbers should challenge the Reserve Bank's view of how "resilient" the labour market, and by extension the broader economy, have proven.
Its February monetary policy statement tells us that 10 of the 14 indicators of labour market conditions it looks at had tightened over the December quarter.
It cited V-shaped rebounds in the NZIER quarterly survey of business opinion's results for firms' hiring intentions and reported difficulty in finding the labour they require.
"Our discussions with business have corroborated this, with firms reporting increased difficulty in hiring both skilled and unskilled labour, particularly in the construction sector," it said. "Closed borders have exacerbated shortages, particularly for workers with specialist skills."
The bank's mandate includes doing what it can to lift employment to its maximum sustainable level. "We currently assess employment as still being below, but closer to, its maximum sustainable level," it concluded.
Surveys and anecdotes are all well and good, but the hard data from the taxman, drawn from what firms file every payday, should be heeded as well. And they suggest a weaker starting point for this year than would have been thought.