New Zealand records a midwife shortfall of 680 full-time equivalents, 20.3 percent of total need, in its 2024 health workforce plan, which projects that the shortage will not grow further by 2033

What is happening?

New Zealand delivers most maternity care through a Lead Maternity Carer, and most of those carers are midwives who are not employed by a hospital. They are paid module by module under a notice made under section 88 of the New Zealand Public Health and Disability Act 2000. The workforce analysis published by Health New Zealand, Te Whatu Ora records about 3,350 midwives in full-time equivalent terms, about half of them employed directly by Health New Zealand in hospitals and facilities and the rest working as self-employed Lead Maternity Carers.

The Health Workforce Plan 2024 records a current shortage of 680 midwives in full-time equivalent terms, which it states as 20.3 percent of total need. The same document estimates no additional shortage by 2033, and attaches a condition to that estimate: training and overseas recruitment have to continue above the current replacement rate. It also records that 38.4 percent of the 187 midwives who graduated in 2023 leave the profession.

The gap appears at the other end as people without a carer. On 2025-06-10 the national professional body for midwives published an analysis of 2022 payment data from the primary maternity services system. That analysis records that 91.5 percent of women registered in pregnancy had a Lead Maternity Carer and 8.2 percent had none at any point during the pregnancy, and that the shortfall concentrates in Auckland, Counties Manukau and the Hutt Valley and falls more heavily on Māori, Pacific, Indian and MELAA families, which is the same set of groups Health New Zealand names on its own maternity programme page when it states that the present system does not work well for Māori, Pacific and disabled families.

Whose problem is this?

The body that counts the shortfall is also the body that commissions the services the shortfall constrains, and the plan it published projects that the shortage will not grow rather than that it will close. Nothing in the material opened here obliges any other body to reach a stated number of midwives by a stated date.

Where does this problem end?

The boundary sits at the supply of carers rather than at the buildings they work in, because a facility with nobody to staff it and a district where no carer is taking bookings produce the same result for the person seeking care. The payment structure is inside the boundary and the level of the payments is outside it, because how a workforce is engaged is a structural question and what it should be paid is a contested one.

What is the state now, and what should it be?

Now

Needs a new measurementthe target state: no source opened here gives a dated official target for eliminating the midwife shortfall. The Health Workforce Plan 2024 states the shortage as 20.3 percent of total need, which implies full staffing against assessed need as the intended end state, and it attaches no date to reaching it. The only figure it gives for 2033 is that the shortage will not grow further, and that figure is conditional on training and overseas recruitment running above the current replacement rate. A date of 2031 appears in a 2023 press report citing a health workforce taskforce report that was not opened, and no source opened here connects that date to the 2024 projection. No source opened here gives a target for the share of women who should have a Lead Maternity Carer, or for the difference between districts or between population groups. This dossier does not supply a target of its own in that empty place.

How big is it?

The shortfall is published as a workforce number rather than as a count of people affected. The Health Workforce Plan 2024 records 680 full-time equivalents short against total need, which it states as 20.3 percent of that need, out of a workforce of about 3,350 full-time equivalents.

The access side is published as a share. In the 2022 primary maternity services payment data, 91.5 percent of women registered in pregnancy had a Lead Maternity Carer and 8.2 percent had none at any point during the pregnancy.

A third figure sits between the two. Of the 187 midwives who graduated in 2023, the workforce plan records that 38.4 percent leave the profession, which is the rate at which the training pipeline loses the people it has just produced.

No count of affected people is derived here. The share of 8.2 percent is measured on the 2022 cohort of women registered for maternity care. The only figure for births that any source opened here gives is a statement on a Health New Zealand programme page that more than 60,000 babies are born each year, and that page attaches no year to it. Multiplying an undated annual total by a share measured in one named year would produce a count whose reference year cannot be stated, so the multiplication is not performed and the population file records the estimate as not derivable.

The two published numbers answer different questions, because one counts the carers missing from the roster and the other counts the share of women who reached the end of a pregnancy without one. Neither of them is a count of the people who would gain a carer if the shortfall closed.

Under what conditions does it arise?

1. The workforce is split between two employment forms and only one of them is counted the same way. About half the midwifery workforce is employed directly by Health New Zealand and the rest work as self-employed Lead Maternity Carers. The same workforce analysis that publishes the 680 figure records that vacancies are harder to track in the community part of the workforce.

2. The training pipeline loses a large share of what it produces. The plan records 187 graduates in 2023 and 38.4 percent of them leaving the profession. Training more people reduces the shortfall only to the extent that the people trained stay in it.

3. The levers pulled so far mostly move money rather than people. The 2020 budget package of NZD 242 million over four years included NZD 85 million to expand the section 88 modules, and the 2024 workforce plan records a shortage of 680 full-time equivalents. No source opened here states what the shortfall was before that package, so the two figures sit beside each other without a stated relationship.

4. The size of the gap is not one number. A press report of 2023-07-06 cites a health workforce taskforce report giving a shortage of 40 percent and a gap closing in 2031. The workforce plan of 2024 gives 20.3 percent and a projection of no additional shortage by 2033. No source opened here sets out the counting basis of either figure.

5. The effect lands district by district before it appears in a national figure. A press report of 2018-02-19 records more than 140 pregnant women in Canterbury unable to find a Lead Maternity Carer over one holiday period, with the same pattern described in Auckland, Tauranga, parts of Waikato and Wanaka. A public broadcaster report of 2025-02-14 records one district left without a resident obstetric specialist, with residents travelling up to three hours each way for specialist care, and a charitably funded birthing centre in another district closing with the plan for a replacement facility on hold.

Each of these conditions is stated in material published by either the health agency itself or by news organisations reporting from the districts affected. None of them is contradicted by anything else opened here, and the first of them is recorded by the agency about its own data.

What has been tried?

Most of the attempts on this record change how the service is paid for or who plans it, and the workforce strategies published in 2024 are the ones aimed directly at the number of midwives. No source opened here reports a measured outcome for any of the five.

What was found?

Why is it still unsolved?

Supply shortage — the model depends on people choosing to enter and stay in self-employed practice, and the only official projection on the record is that their number holds rather than rises to meet assessed need.

The shortfall is a count of people, and most of what has been done moves money. A notice sets what each module of care pays. A package in 2020 put NZD 85 million into those modules. A commissioning framework changes who plans the service in each district. None of those steps changes how many people finish midwifery training and remain in practice, and the plan that publishes the shortfall also publishes that 38.4 percent of the 2023 graduating cohort leave the profession.

The second part is that about half of the shortage has no meter. Roughly half the workforce is employed directly and the rest is self-employed, and the same analysis that gives the 680 figure records that vacancies are harder to track on the community side. An unfilled hospital position is a line in a staffing return. A district where no self-employed carer is taking new bookings produces no vacancy anywhere, and the person who cannot find a carer enters the record only later, as a share of a payment dataset published by a professional body.

The third part is that the size of the gap is given two different values and the horizon has moved without a stated link between the two. One lineage puts the shortage at 40 percent and describes the gap closing in 2031. The workforce plan of 2024 puts it at 20.3 percent and projects that it will not grow further by 2033, on the condition that training and overseas recruitment run above the replacement rate. No source opened here states the counting basis of either figure or connects the two dates. A shortfall carrying two values with no published reconciliation cannot produce a single missed target.

The fourth part, which is the one that keeps the pattern stable, is that the failure produces no event. There is no dated public target for eliminating the shortfall, so no year arrives in which one is missed. The share of women without a carer is published by a professional body rather than as an official indicator with a threshold attached. Reports of a district losing a service arrive one district at a time, and each of them reads locally rather than as an instalment of a national count.

What observation would mean it is solved?

Candidates — (a) Health New Zealand publishes a workforce update recording the midwife shortfall at or near zero against its own assessed need, with the date on which it was measured (b) the share of women registered in pregnancy who have a Lead Maternity Carer rises toward the whole cohort in the payment data, and the difference between districts and between population groups narrows (c) the share of each graduating cohort that remains in practice rises across several consecutive years.

(a) alone counts the roster and not the care. A national total can reach assessed need while Auckland, Counties Manukau and the Hutt Valley stay short, because the shortfall is published nationally and the access gap recorded in the 2022 payment data is regional. Assessed need is also a figure the same body sets, so the gap between supply and need can close from either side of the subtraction.

(b) alone can move for reasons outside the workforce. The measure is built from payment records for women who registered, so a change in who registers changes the denominator. The 8.2 percent is one year of one dataset, and no source opened here publishes the same measure for any other year, so there is no series from which a direction could be read.

(c) alone is upstream and slow. Retention sets how much of each cohort survives into practice, and a cohort of 187 is small against a workforce of about 3,350 full-time equivalents. Better retention would take years to appear as a smaller shortfall, and it says nothing about which districts the people who stay end up working in. The three have to be read together, and (b) is the one that speaks to the person who could not find a carer.

What is it connected to?

Fills with researchthe obstetric and anaesthetic workforce in the same facilities, the viability of rural hospital maternity services, recognition pathways for overseas-qualified health workers, housing and living costs as drivers of where health workers settle, primary care access for pregnant women outside the Lead Maternity Carer system, and maternity workforce arrangements in Australia and other comparable jurisdictions. Relation type and evidence grade were not confirmed in this round.

What these sources do not say

• Why the two shortage figures differ. One lineage gives 40 percent and the workforce plan gives 20.3 percent for a broadly similar period. No source opened here sets out the counting basis of either, and neither acknowledges the other.
• How the 2031 date and the 2033 projection relate. A 2023 press report describes the gap closing in 2031 and the 2024 workforce plan projects no additional shortage by 2033. Whether those are one calculation updated or two calculations resting on different premises is not stated in anything opened here.
• What the 8.2 percent converts to in people. The count of births or registrations for 2022 was not obtained. The official statistics release for births and deaths for the year ended December 2022 returned only its title to the automated reader, so the denominator is not stated here and no count of affected people is derived.
• The text of section 88 itself. Only the gazetted notice made under it was opened. The legislation site did not return the section to the automated reader, so the statutory wording is neither reproduced nor characterised here.
• What self-employed Lead Maternity Carers actually earn. The module rates sit in the notice. No source opened here publishes the distribution of actual income, or the effective hourly return, for carers working under it.
• How many community vacancies there are. Health New Zealand records that vacancies are harder to track in the community part of the workforce, and no source opened here publishes a count for that part. The absence of that count is recorded by the agency itself.
• The vacancy rate in the employed workforce. A news report exists whose subject is the midwife vacancy rate in the Health New Zealand workforce, and its text was not received by the automated reader, so no vacancy figure is stated here.
• A national picture of regional access. The districts named in the 2022 analysis are the districts that analysis names. No map or table covering every district was found, so nothing here says how the access gap is distributed outside those districts.
• Whether the 2025 recommendations were adopted. No source opened here reports a government response to them.
• What Kahu Taurima has changed. The framework begins regionally in 2025 with national completion targeted for 2029-07-01. Nothing opened here reports a measured effect on access.
• A response to these findings from the bodies named. The only position found is a written reply reported in 2023, stating an intention to deploy the workforce equitably and sustainably. No statement, rebuttal or explanation from Health New Zealand or the Ministry of Health addressing the 2024 shortfall figure or the 2022 access gap was found in anything opened here.

See the evidence

No primary workforce plan document was read in full. The Health New Zealand midwifery workforce analysis page and the maternity programme page were opened directly, as was the gazetted section 88 notice; together they carry the 680 figure and its 20.3 percent statement, the conditional projection to 2033, the graduate attrition rate, the workforce size and split, the agency record that community vacancies are harder to track, the annual birth statement, the Kahu Taurima timetable and the payment structure. The access figures of 91.5 percent and 8.2 percent come from an analysis of payment data published by a professional body, and the payment data behind them was not opened. The competing shortage figure of 40 percent and the 2031 date come from a press report citing a taskforce report that was not opened, which is why they are graded low and appear in the body as a second figure rather than as the figure. The regional service losses and the earlier access failures come from a public broadcaster report and a newspaper report respectively. Three rows are carried without a URL rather than removed: the ministry publication page returned HTTP 403, and the official births release and the vacancy report returned titles without bodies, so no birth count and no vacancy rate appear anywhere in this dossier. Where the opened sources disagree the disagreement is left visible rather than resolved, and the disagreement over the size of the shortage is treated in the body as part of the problem. This is a Path A output (research-based definition), so observation_refs is empty and provenance_mode: press-derived.

People affected

Estimated range Not derivable

The reason and what is missing are listed under “What is missing” below