I have been arguing for pro-natalist policy in this country for years. So let me begin with credit where it is due. New Zealand First has this week put family formation into the election discussion.
The Kiwi Kids Grant would pay $5,000 a year, tax free, for the first three years of a child’s life. It would apply to each of a family’s first three children provided at least one parent is a New Zealand citizen. It would be paid monthly through Inland Revenue, with a lump-sum option in year one for first-time parents facing the cot-pram-car-seat nightmare.
It is a good instinct and I would like to see something like it become government policy.
But the case being made for it is the wrong case, the design is not optimised for the scarce dollars the government has and the costing is out by about a quarter.
The numbers do not hold
Here are some of the problems with NZ First’s announcement, which show a level of slipshodderiness which is not surprising but which is still frustrating.
The decade that is nineteen years
The release says citizen births fell “over the past decade from 52,506 in 2006 to just 36,351 in 2025. That is a 30% drop in just the last ten years.” That is nineteen years. The percentage is roughly right and the trend is real, but the timeframe is off by nearly a decade, and 2006 is a peculiar place to start. Stats NZ’s own guidance records a baby blip from 2007 to 2012 that took the total fertility rate to nearly 2.2 before it fell steadily to under 1.6 by 2025.
The base year therefore sits at the foot of the strongest fertility run New Zealand had seen in thirty years. Our fertility numbers are bad enough without being stated in a way that hands anti-natalists a free hit.
The births that are not there
Add NZ First’s two figures together: 36,351 citizen births plus 14,380 non-citizen births gets you 50,731 births in total. Stats NZ registered 57,705 live births in 2025.
That leaves roughly 7,000 babies, about one in eight, unaccounted for. Which means the “72 per cent citizen, 28 per cent non-citizen” split is being calculated on a base missing an eighth of all the children born in this country.
No source is cited by NZ First, so I cannot check the working. The figures plainly use different populations or definitions. Until the party says what they are, the claimed 72/28 split cannot safely be treated as a breakdown of all births.
The costing is too optimistic
The release puts the cost at “$400 million by the third year once the full cycle is established.” There was some confusion about what this means. RNZ reported it as $400 million over three years. The Spinoff read it as $400 million a year. For what it’s worth the natural reading is that the scheme would cost $400 million a year by its third year, once three cohorts were receiving it.
But the arithmetic does not work either way. On New Zealand First’s own birth figure, stripping out fourth and subsequent children as the policy requires (about nine per cent of births), roughly 33,000 children enter each year. At steady state you are paying three cohorts at once, about 99,000 children at $5,000 each, which is $496 million a year.
On the party’s own apparent assumptions, then, the costing looks to be short by about a quarter. And of course the precise figure cannot be known until NZ First discloses how it calculated the eligible population.
Giving with one hand and taking with another
New Zealand First wants to bolt its policy onto the Best Start program, which pays $77.00 a week, or up to $4,041.00 a year for every child under three. Since April this year, Best Start has been income-tested from year one, abating at 21 per cent above $79,000 and cutting out entirely just over $98,200.
That was a change made by the government New Zealand First is part of. The party now proposes a second payment for the same children over the same three years, universal where the first is abating, so that identical babies attract two overlapping entitlements on two different tests, both administered by Inland Revenue.
The terminology is bad
I would let pass if it were the only thing. The release twice gives the fertility rate as births “per person.” But the measure of a total fertility rate (TFR) is births per woman.
This is not pedantry. Well, it’s not just pedantry. The substitution makes both sentences arithmetically meaningless: divide last year’s births by every person in the country and you get something nearer 0.011, while a replacement rate of 2.1 per person would describe a population doubling every generation rather than holding steady.
What this shows is that nobody who understood the statistic read the document before it went out. That matters less for what it says about the fertility rate than for what it implies about everything else in the policy. A party that cannot state its headline measure correctly is not one whose unsourced citizenship figures I am inclined to take on trust.
And there is an irony here.
Ten weeks ago New Zealand First’s own member’s bill passed its first reading, 67 votes to 55, on the proposition that “woman” should mean “an adult human biological female” throughout the statute book, because, in Winston Peters’s words earlier this year, the country ought to “focus on the facts of biology.” One of the facts of biology is that babies are born to women, which is why the fertility statistic is expressed per woman rather than per person (at least according to NZ First).
The overseas evidence is decidedly mixed
The received wisdom on cash-for-babies is that it shifts the timing of births without changing overall family size. However, the evidence has never been entirely one-sided. Recent Australian work makes it much harder to maintain.
Back in 2004, Australia implemented a $3,000 AUD baby bonus. This was increased to $4,000 and then $5,000 before the payment for second and subsequent children was reduced in 2013. The scheme was abolished the following year.
The e61 Institute’s October 2025 work on the policy, which used administrative data and sharp month-to-month cut-offs nine months after each policy announcement, found that the 2004 scheme lifted births by around six and a half per cent. The affected cohort had 6.8 per cent more children by 2022, including an increase in the number of women who became mothers.
The effects were largest among older mothers and third-and-higher births. These are precisely the cohorts where a mere timing shift is least plausible. And e61 found no evidence of compressed birth intervals.
In Quebec, the Allowance for Newborn Children raised fertility by about 12 per cent overall, and by around 25 per cent among families eligible for the full $8,000 CAD payment, which was reserved for third and subsequent births.
That said, we shouldn’t overstate the case.
In Poland, a program known as the Family 500+ is even more generous than what Australia offered. The policy paid 500 złoty a month, tax-free, for every eligible child up to the age of eighteen. This was close to 40 per cent of the net minimum wage at the time. The payment was raised to 800 złoty in 2024.
However, looking closely, the effect was only 0.7 to 1.8 points among women aged 31 to 40 while reducing it by 2.2 to 2.6 points among women aged 21 to 30.
One plausible reading, although the study cannot prove it, is that older women brought forward births they had already planned while younger women deferred them. And of course at no point did Poland’s TFR threaten to rebound to replacement.
If there’s any consistent way to read all this it’s the following:
Universality mostly moves money.
Payments scaled to birth order, and payments that are simply generous, do move fertility.
But cash payments do not move fertility enough to reverse the decline.
There may also be a labour-supply cost. Separate work on Poland suggests that some of the transfer reduced maternal employment and earnings.
That does not bother me because supporting parents to spend more time raising children is not necessarily a policy failure. Caring for a small child is work. It is simply work the national accounts have no way of seeing. A mother who drops from five days to three has not become less productive. She has shifted some of her production into the one sector we decline to measure.
If a transfer designed to defray the cost of raising children ends up letting parents buy back some of their own time in order to raise them, the transfer has done precisely what it was meant to do. But others feel differently.
What the Kiwi Kids Grant would actually buy
Let’s use the Australian estimate as an optimistic illustration and apply it to a notional eligible New Zealand cohort of about 33,000:
A 6.5 per cent lift is roughly 2,150 additional babies a year, at about $246,000 each.
If we assume, for illustration, that 63 per cent of births are eligible, that works through to a total fertility rate of about 1.61, against 1.55 now.
At the low end of the international range, two per cent, you get about 660 extra babies at roughly $765,000 each and a TFR of 1.57.
The gap between where we are and replacement is 0.55 of a child per woman. So this policy, on the most encouraging evidence available anywhere, closes about a ninth of it. On cautious assumptions, a thirtieth.
Meanwhile between 94 and 98 cents in every dollar goes to a child who was going to be born anyway.
I want a policy like this to exist, which is exactly why I will not pretend the elasticity is bigger than it is. Sell a half-billion-dollar program on a promise to reverse fertility decline and you have handed your opponents a five-year countdown to a graph showing the birth rate still falling.
The argument that actually wins
The 94 cents is only waste if you believe the point of the policy is to change behaviour. It is not, or it should not be. That is not what government is for.
New Zealand Superannuation is universal, unfunded and paid out of current taxation. It is not a savings scheme. It is a transfer from working-age people to seniors, and we are well past the point where retirees draw more from the system than they put into it.
Every child raised in this country is a future contributor to it. Yet every child raised in this country is also, overwhelmingly, paid for privately, by parents, out of their after-tax income. That cost that runs comfortably past a quarter of a million dollars out of pocket, before you even begin counting foregone earnings.
Consider two people who retire on the same day on the same pension:
