In November 2018, officials at Immigration New Zealand began work on replacing the ageing biometric systems that verify identities and manage immigration risk. Seven years later it was dead, having delivered no measurable benefit. Nobody can yet say what it cost.
The Ministry of Business, Innovation and Employment (MBIE), which houses Immigration New Zealand, told Parliament in June that around $33 million would be written off. On 29 July, Immigration Minister Erica Stanford disclosed that the ministry’s chief executive had found a further $6 million in associated costs. He could not assure her that even this was the full amount.
An independent financial audit is under way because the ministry cannot establish the figure itself.
The ministry had earlier commissioned an outside reviewer, Greg James, whose report, released in June, makes for extraordinary reading. The project began without ministerial sign-off and grew into a major investment before any effective authority caught up with it. Its business case called the project cost-neutral and self-funding, projecting $18.79 million in savings over five years that no analysis supported.
Officials revised the whole-of-life cost upward from $19.5 million in 2019 until it pressed against $35 million, the threshold above which Cabinet approval becomes mandatory.
In 2024 they sought an increase to $39.9 million through an unrelated Cabinet paper on fees and levies. Stanford refused and asked what the money was buying. The answer was a restructuring: costs moved between workstreams, $4.44 million moved out of the project altogether, and the reported figure settled below the threshold again. The review records staff describing it as ‘creative accounting’, driven largely by efforts to stay under the Cabinet line.
Officials had already taken the same request to Andrew Little, Stanford’s Labour predecessor in the immigration portfolio, and been turned down, twice by her account. When they came to her, they did not mention his refusals.
Stanford told Parliament that officials deliberately withheld information from her and from the previous government alike. The review itself stopped short of finding deliberate intent, and Michael Heron KC is now investigating whether the minister’s harsher reading is the correct one. Even the review sat inside the ministry for two months before anyone showed it to her.
Australian readers will recognise the genre. The ABC’s Utopia has spent a decade documenting the fictional Nation Building Authority: the project boards, the reset workshops, the rebranding exercises that substitute for delivery. One thing even the satirists never dared to show was a department giving up on delivery altogether and putting its effort into appearances instead.
Most commentary in Wellington has treated the affair as a failure of project management. And the review gives that reading plenty to work with: unwarranted confidence, weak governance, ignored assurance. Yet bad project management does not explain the detail that should worry us most. Officials gave misleading advice to two governments in succession.
A bureaucracy misleading only one government might be explained by politicisation. But this one misled both a centre-left and a centre-right government, which points to its loyalty running somewhere other than towards the government of the day. A century ago, the German sociologist Max Weber gave the condition a name: Beamtenherrschaft, the rule of officials.
I run a think tank in Wellington, a short walk from most ministries. Much of what I pick up about the public service cannot be published without identifying people who spoke in confidence. And I am certainly not the only person in Wellington who would be aware of the problems in the public service without the ability to reveal them.
The biometric case therefore matters because this time the behaviour is on the public record. An independent reviewer, a select committee, a Privileges Committee and now a King’s Counsel have all been over it.
Of course, there are outstanding public servants in New Zealand, people of intelligence and dedication who joined to serve. But the system stopped rewarding outstanding service a long time ago. To understand what it rewards instead, I went back to Max Weber.
Weber was compulsory reading when I studied economics and law in Germany. Still, I did not expect to need him most in Wellington.
Weber delivered his famous lecture on politics as a vocation in January 1919, in a Munich gripped by revolution. His portrait of the modern official remains the best yardstick for judging a public service.
The genuine official, Weber argued, administers sine ira et studio, without anger or partiality. He gives his political masters frank advice, and when a decision goes against that advice, he executes it as conscientiously as if it matched his own conviction. Weber considered this discipline a form of moral honour, not servility.
Two features made such an official possible. He was a specialist, a Fachbeamter whose authority rested on mastery of his field, acquired over long years of training and work in the same field. And he held an office, an Amt, which carried an ethos. The office was a vocation rather than a stepping stone to power, and its holder answered to its standards even when nobody was watching.
Weber was not an enemy of bureaucracy as such. He believed trained officialdom to be technically superior to any amateur administration.
Officials command what Weber called Fachwissen and Dienstwissen, specialist knowledge and the ‘knowledge of the files’. A minister depends on what officials choose to show him. Beyond political control, he warned, such an officialdom would end up ruling without answering for anything.
For the opposite of his specialist official, Weber looked to China.
In his study of Chinese religion, Weber examined the literati, the mandarin class whose examinations selected China’s governing elite for thirteen centuries until 1905. The system was meritocratic, at least beside hereditary patronage. It gave a vast empire a common written culture, but the merit it rewarded was of a particular kind.
The examinations required mastery of the Confucian canon and its approved literary forms. The curriculum shifted over the centuries and included current policy questions, but the centre of the test always remained the same. By the Ming dynasty, it had hardened into the eight-legged essay, a rigid form arranging canonical material into eight prescribed sections. The system reliably measured whether a man wrote like his examiners, not whether he could manage a flood or a famine.
Once appointed, the mandarin stayed a generalist by design. Under the rule of avoidance, he could not serve in his home province. The empire rotated him to a new posting every few years, so that he would never develop local attachments. The price was that he never acquired local knowledge, either.
Much of the daily work fell to local clerks who stayed put while their superiors circulated above them. The clerks accumulated the knowledge and much of the informal power. The rotating magistrate carried the formal responsibility.
The mandarinate recruited its own successors, sat in judgement on its own performance and set the standards by which both were measured. Its loyalty ran to the class and its canon rather than to any particular emperor’s programme, and it survived changes of dynasty that swept away nearly everything else.
For Weber they were the opposite of his modern official: cultivated generalists rather than trained specialists. They were a status group bound by canon and honour rather than by service to any programme.
Westminster should remember this, because its own civil service owes one branch of its ancestry to the mandarinate.
In June 1853 the Commons debated recruitment for India. Lord Stanley noted that unlimited competition for public office was said to prevail in China, ‘and therefore it might be called the Chinese principle’. He thought it infinitely preferable to patronage. His one worry was that, applied rigidly, it risked flooding India with ‘over-educated mediocrity’.
The same principle soon reached the home civil service. Sir Stafford Northcote, a future Chancellor of the Exchequer, and Sir Charles Trevelyan, the Treasury’s most senior official, wrote it into their report, published in 1854. In time, senior British officials became ‘the mandarins’.
Britain drew much of its administrative elite from classically educated generalists, men examined in Greek verse and sent to run departments whose work they had never studied. By 1968 the Fulton Report was attacking the civil service’s ‘philosophy of the amateur’.
But two things were meant to keep the Westminster mandarin in check. Permanence meant a career inside the service, often deep inside a single field, from which a generalist entrant could still emerge, decades later, as a genuine expert. The ethos of office made serving the elected government of the day the point of a bureaucratic career.
The British model was then exported around the Empire, including to New Zealand. In 1988, New Zealand radically changed it.
The State Sector Act belonged to one of the most radical market-liberal reform programmes any democracy has attempted. The reformers around Finance Minister Roger Douglas privatised, deregulated and liberalised at a speed that made New Zealand a pilgrimage site for free-market economists. Much of it was overdue.
The public service did not escape the sweeping reforms. The 1988 Act abolished permanent department heads and replaced them with chief executives on fixed-term contracts. Ministers would purchase outputs from their departments. Managers, in the slogan of the day, would be left to manage. Incentives would do the rest.
As theory, it was impeccable. Agency theory taught that officials, like everyone else, respond to incentives, so the answer was to write better contracts. Managerialism, imported alongside the economics, treated management as a transferable skill, so leaders could move between portfolios as freely as executives move between companies.
Foreign officials came to study what became known as the New Zealand model. The reformers believed they were finally abolishing the mandarin, the complacent permanent head who answered to nobody and outlasted every government.
But as it turns out, they abolished the wrong things. Fixed-term contracts at the top weakened the link between long service in a field and the job of running it. Generic management outranked subject knowledge in the model’s scale of values. What survived was what Weber had described in China: the generalist career, the perpetual rotation and the examination in style rather than substance. Managers left to manage had become mandarins.
The sharpest critics were not outsiders. Reviewing the reforms in 1996, the American budget scholar Allen Schick warned that contract-style accountability could harden into a ‘checklist mentality’, since responsibility rests on judgement as much as on outputs. In 2019, Simon Chapple of Victoria University measured the result. Departmental chief executives and their senior teams were averaging barely two years in their posts, in what he called a drift from mandarin to valet.
He was half right. The service Chapple describes is a valet in its manners and a mandarin in its grip on information. That is a dangerous combination.
Today’s senior official moves between agencies every two or three years, treating rotation as a career strategy. Job advertisements ask for comfort with ambiguity rather than knowledge of anything in particular.
In 2022, the Treasury advertised for a senior analyst in its economic strategy unit and noted that an economics background was ‘not essential’. The competency framework is New Zealand’s version of the eight-legged essay: a canon of approved behaviours unrelated to the job.
The Public Service Commissioner, himself an official, runs the appointment of every departmental chief executive and remains their statutory employer. Ministers are consulted, and Cabinet may reject a recommendation. But a minister does not ordinarily hire or dismiss the person who runs the department for which he answers to Parliament. And so, senior officials move constantly between agencies, and whoever wins each election finds much the same people running the departments.
Wellington’s geography does not help. Ministers work in the Beehive while their departments occupy towers scattered across the city. A minister learns what the ministries are doing the way an emperor learned about a distant province: through papers (even when the departments are only a few hundred metres away).
A ‘spirit of service’ has sat in New Zealand’s public service statutes since 1962. It survived the 1988 reforms, and the 2020 Public Service Act declared it the service’s ‘fundamental characteristic’. Parliament repealed that declaration in June this year, although the phrase survives in the Act’s purpose. Quoting that spirit has become a form of self-reassurance for the public service, even though it has become less clear what it means in practice.
A self-recruiting class also develops shared instincts, just as the literati had Confucianism. It acquires a settled sense of which options are responsible, which risks are tolerable and which proposals need never reach the minister. The complaint I hear from ministers, from more than one party, is not that officials carry party cards. It is that advice arrives pre-shaped, options already narrowed, and that the department presents its inherited view as the neutral baseline.
None of this requires villains. Officials respond to the incentives in front of them, which is what economics expects everyone to do.
A public servant takes the blame for the visible error: the consented project that fails, the approval that generates headlines. Nobody takes blame for the reform not attempted, the project quietly not delivered, the years lost to process. A system that punishes visible error and ignores invisible failure will select, promotion round after promotion round, for exactly the caution, slowness and risk aversion for which Wellington is now known. The ‘no surprises’ doctrine, found in no other Westminster system, obliges officials to warn ministers of anything embarrassing.
The biometric project I mentioned at the beginning ran on that mechanism from the start. At each decision point, silence was a safer personal choice. Admitting failure created an identifiable loser, while carrying on spread the cost and accountability across budgets, years and successors.
Weber’s ethos of office existed to defeat that private calculus, to make candour a duty of the Amt rather than a career risk. But the 1988 reforms assumed the ethos would survive on its own while contracts did the enforcing. They removed everything that had sustained it.
In his 1997 book The Audit Society, the British accounting scholar Michael Power warned that audit breeds its own rituals. Organisations learn to satisfy those rituals without improving anything underneath. The biometric project satisfied them for years. Managerial accountability turned out to be a set of documents managers could produce.
The response to the immigration scandal has so far followed the usual pattern. The chief executive apologised to Parliament, promised to learn the lessons and announced a stocktake and stronger processes.
No dismissal has been announced at the time of writing. If the KC’s findings warrant any, the decision will rest with the relevant employer, which is to say with the public service itself. Ministers rarely force the issue either, because programmes never evaluated embarrass nobody. The Confucian bureaucracy handled the emperor’s censors much the same way. Everyone was heard, everything was recorded, and nothing changed.
Reform in New Zealand increasingly happens despite the ministries rather than through them. When Erica Stanford, as Education Minister, overturned decades of failed literacy teaching, she took her advice from an external ministerial advisory group. Its chair, I should disclose, was a senior fellow of my own think tank. The same minister who went outside one ministry (Education) for advice was misled by another (MBIE).
Weber’s own country draws the line differently. Germany places State Secretaries at the apex of every ministry. Administrative State Secretaries are the ministry’s most senior civil servants. The government can move them into temporary retirement the moment ministerial confidence is lost. Below that level, career officials enjoy strong tenure and a statutory duty to object to unlawful instructions.
The arrangement predates Weber and is no Weberian paradise. German scholars describe a senior service deeply engaged in politics at the leadership level. But its virtue is its transparency. The law says openly where political confidence matters. That honesty protects the officials beneath it better than New Zealand’s myth of neutrality protects anyone.
The New Zealand reformers of 1988 read their principal-agent theory closely. Weber they did not reject. They just took him for granted.
So, the question for New Zealand’s public service today is whether a mandarinate, once entrenched, can reform itself at all. The evidence from the original Chinese mandarinate suggests that it cannot.
The examination elite had every interest in preserving the institution that conferred its status. Some of its members pressed for change, but the system outlasted centuries of proposals to mend it.
The court abolished the examinations by edict in 1905, after decades of defeat and Japan’s modernising example had made reform unavoidable. But it came too late to matter and too suddenly to manage. The old route to elite status vanished before anything stood ready to replace it.
Economists later found that the prefectures with the most examination places per head supplied the most revolutionaries afterwards and ran the highest risk of uprising in 1911. A route to status had closed, and disappointed aspirants turned on the state. The dynasty they had served for so long, and slowly strangled, fell six years later.
Mandarinates rarely reform themselves from within. They are eventually overtaken by reality, and the costs arrive in two instalments: one for the years of drift and a larger one for the demolition when it finally comes.
New Zealand’s instalments are already arriving: a biometric project here, a generation taught to read by discredited methods there, a health system busy rationing care.
Meanwhile the economy sits near the bottom of the OECD for output per hour worked, and the officials who might change that keep rotating onward, as comfortable as ever with ambiguity.
Seven years of work bought New Zealand no new biometric system. Nobody can yet say what that failure has cost.
But the files bought something almost as valuable: an X-ray of the state. The image shows a class that appoints its own members, judges its own performance and, when the accounting arrives, answers only to itself.
It took defeat in multiple wars and internal rebellions for the Qing to change course. A democracy is supposed to accomplish that through elections.
But if the biometric files are anything to go by, it is not clear that New Zealand elections can affect the dominance of the public service at all.
To read the article on the Quadrant website, click here.