Unemployment numbers that don't stack up

The unemployment rate has dropped to 3.9% - a great result for the government.

But since the non-publication of latest official child poverty data due to "uncertainty" and a "lack of confidence" in the Statistics NZ  Household Economic Survey sampling, I am wary. More wary than I was anyway.

The unemployment data comes from the Statistics NZ  Household Labour Force Survey.

I had a dig into the tables looking for any stand out development.

Here's one.

In the Manawatu-Wanganui region, the unemployment rate (2nd to last column above) between June and Sept 2018  dropped three whole points from 6.6 to 3.6 percent.

This should be reflected in benefit statistics, no?

It isn't. The number on Jobseeker Support rose.

I checked out the number for the Manawatu-Wanganui region - a different stat which slightly more closely matches the region surveyed in the HLFS.

In June 2018 there were 8,352 people on a Jobseeker benefit: in Sept 2018, 8,532.

The Taxpayer's Union has also questioned the broader opposing trends.

We can measure unemployment three ways: through the HLFS, through the numbers on unemployment benefit and via the Census. Obviously the last count is too infrequent and time-lags terribly.

Just be aware that the positive HLFS result is not mirrored in the benefit data result.

The HLFS result is probably a facet of the growing working age population and labour force. The denominator is increasing faster than the numerator. But it could also be a 'rogue' result.