Wednesday, May 26, 2021

Lindsay Mitchell: Can Government Numbers be Trusted?

Occasionally I get told off for writing about 'child poverty'. People say I shouldn't use the term because real child poverty doesn't exist in New Zealand. Or that there is no such thing as child poverty - only parental poverty. They are fair enough criticisms given the World Bank defines extreme (or absolute) poverty as living on $1.95 a day or less. New Zealand only has relative poverty yet too often the descriptor is missing.

But the language must be used to dispute the subject. There's not a lot to be gained from stuffing our fingers in our ears when so much this majority government does revolves around reducing child poverty.

Let's try and simplify how child poverty is defined and measured without too much jargon.

The survey which produces the numbers is the Household Economic Survey (HES) run by Statistics New Zealand. The sample was increased from around 3,500 to 28,500 dwellings to provide more robust data.

Respondents report income received from all sources to a visiting interviewer. By law, no-one can refuse to participate (though clearly some do as the survey targets 28,500 to ensure 20,000 responses!) I was a two-year participant in a stats NZ survey back in the early-2000s and can tell you it is a pain in the neck. Which brings me to my first point about the quality of the data.

Being short on time, having difficulty with English, feeling resentful about the current political climate are just a few reasons why one might go through the motions, make up some answers to satisfy the interviewer. Some may even lie if the family illegally receives benefit income as well as income from work. On what basis would I say that? An Auckland University study using Growing Up in New Zealand data found, “70% of those who say they receive the domestic-purposes benefit also answer yes to the question of whether they have a partner.” Respondents certainly aren't going to declare ‘cashies’ or income from selling illegal substances.

Some might simply think their income is none of the government’s damn business.

Participants are told they will not be paid but they must supply information for the good of the country in the same way they pay taxes for the good of the country.[i] Statisticians obviously believe a sense of civic duty abounds among the citizens.

Anyway, once Statistics NZ identifies the household income it's then put through a process of 'equivalisation'. This is a downward scaling of actual income adjusted for the number of household occupants.

A household with 2 adults and 4 children with a total income of $94,500 would be equivalised to $35,000; a household with 2 adults and 2 children with the same total income would be equivalised to $45,000.

Thus, a family with four children may be reported as poorer than the family with two children living next door even though their total incomes are exactly the same.

Once a median equivalised household income is established, various thresholds are drawn at 40, 50, and 60 percent. Children living in the households under the 40% threshold are the poorest.

‘Before housing costs' and 'after housing costs' incomes are differentiated to add another dimension.

These measures form the basis of the relative child poverty stats.

Obviously, there are problems with relativity measures.

Recently the data had to be adjusted because Treasury, "...identified several respondents incorrectly reporting the superannuation payments they received, resulting in double counting their income from superannuation." This resulted in a change to the median income for the year ended June 2020 which reduced from $42,486 to $41,472. Consequently, fewer children fall below the re-calibrated threshold. Fewer children are in poverty. But nothing materially changes for those children.  As the Children's Commissioner noted, "A different number behind a decimal point doesn’t change things for the thousands of tamariki and whanau doing it tough."


In Ireland in the 1990s, using relative measures, “the poor became ‘richer’ in real terms, but because the income growth of the middle income households was even greater, poverty rates grew considerably.”[ii]

Conversely, if households get collectively poorer in a recession, fewer children will fall below the thresholds leading to an apparent improvement in child poverty. I'm sure that isn't Labour's intention. To make us all poorer to claim a reduction in child poverty - the PM's flagship policy. Then again, she shows no alarm at the prospect of more children being on benefits as long as she can claim to have lifted them out of poverty.

Speaking of benefits much has been made of the decrease in the unemployment rate.

That number is also the product of a survey, this one the Household Labour Force Survey (HLFS) of 15,000 individuals.

Some observers have noticed that the unemployment rate is 4.7% whereas 6.3% of working age people are on a Jobseeker benefit. Why? Two reasons (there are more).

One: the HLFS covers individuals aged 15+ whereas the Jobseeker benefit covers 18–64-year-olds. It beats me why the HLFS persists in surveying 15-year-olds who aren’t even legally allowed to leave school. Anyway, the bigger denominator provides a smaller (better) fraction.

Two: people can be on a Jobseeker benefit AND working part-time. From April 1, 2021 beneficiaries can earn up to $160 before their benefit is reduced. This is a large rise from the previous $90 on Jobseeker and $115 on Sole Parent Support. It is probable that more people will become part workers/part beneficiaries (which will look good for the unemployment rate but bad for benefit numbers.)

You can guess which measure the government will trumpet.




Lindsay Mitchell is a welfare commentator who blogs HERE.

1 comment:

DeeM said...

What's the famous quote..."there are lies, damned lies and statistics".
You can extract pretty much any figure you want from a body of data if you are deliberately subjective and can come up with a justification to "filter" the numbers you don't like.
That's how a lot of "Science" operates these days, in particular research into human-induced climate change theory, or fact as our governments would like us to believe. There is a large body of actual research empirical data which casts doubt or contradicts this theory and yet we persist in clinging to it like a drowning man to a lifebelt.
Government policy is rife with "statistical enhancement" to fit their policies and produce the "right result".
All governments do it, and these days many scientists as well, which shows the sad state of affairs we've gotten ourselves into.