The Hidden Architecture of Government Failure

The Long Read

Bad incentives, not bad people, keep breaking New Zealand’s institutions – and only structural change will fix them.

New Zealand has never spent more on public services, administered by a workforce that has grown by a third in less than a decade. And yet hospital waiting lists stretch into years, students leave school unable to read, and infrastructure crumbles.

More money. More people. Worse results. If resources are not the problem, what is?

The answer lies in the incentives built into government’s architecture. That architecture determines whether those who make decisions also bear the consequences. When the two are aligned, systems work. When they drift apart, failure becomes inevitable – not because people stop caring, but because the system makes success heroic rather than routine.

Everyone understands these incentives intuitively. In a flat where the heating bill is split evenly, everyone runs the heat a little higher than they would if paying alone. No one is selfish – but each flatmate has authority over the thermostat while bearing only a fraction of the cost. Multiply that logic across a $183 billion government and you begin to see the problem.

New Zealand has built a state where authority and consequences have been systematically separated. Decisions are made in one place. Costs fall in another. Outcomes are unmeasured or ignored. Little wonder the government’s heating bill has exploded while other needs go unmet. This is not a story about villains. It is a story about design.

Why Good People Produce Bad Outcomes

A generation of the world’s most influential economists has mapped this problem. Their work earned Nobel Prizes and reshaped our understanding of how governments and firms work. Their conclusion was simple but uncomfortable: when authority is separated from consequences, failure follows – regardless of intentions.

James Buchanan, founder of public choice theory, warned in Public Choice: Politics Without Romance (1979) that public servants respond to the same pressures and incentives as everyone else. Systems built on the assumption of exceptional virtue, he wrote, were “destined to disappoint.” William Niskanen showed why: agencies tend to maximise their budgets because budgets determine staffing, prestige, and security. The tendency is structural, not moral.

Bengt Holmström, later a Nobel Laureate, showed in Moral Hazard and Observability (1979) that behaviour changes when decision-makers don’t bear the consequences of their actions. When responsibility sits in one place, information in another, and consequences somewhere else again, systems generate blind spots. Good people cannot bridge structural misalignment.

Oliver Hart, awarded the Nobel Prize in 2016, clarified how organisational performance depends on who holds decision rights. When those with formal authority don’t share in the gains or losses from their decisions, caution and silence dominate. Elinor Ostrom offered the reverse insight: in Governing the Commons (1990), she showed that systems thrive when authority, information, and accountability sit close together.

These thinkers converged on a single principle: authority without consequences produces predictable failure. Separate power from accountability, and the system will drift toward dysfunction. New Zealand’s public institutions exhibit exactly the conditions they warned about.

How the Architecture Fails

The failures of education, health, housing, and regulation look different on the surface. Underneath, they share the same flaw: authority in one place, consequences in another.

Education operated for two decades on constructivism – the idea that children learn best by discovering knowledge for themselves rather than being explicitly taught. The “numeracy project” emphasised strategies over basic mathematical knowledge. Whole-language approaches displaced systematic phonics.

The consequences were predictable to anyone familiar with cognitive science. Reading, writing, and mathematics require explicit, systematic instruction. The brain is not wired to infer phonics or discover long division. When curricula de-emphasised direct teaching, outcomes collapsed.

Since PISA testing began in 2000, New Zealand’s mathematics score has fallen 44 points – equivalent to one-and-a-half years of schooling. In reading, we have dropped 28 points. High performers in maths have halved; low performers have doubled. The children who suffered most were those whose homes could not compensate for what schools failed to teach.

Authority over curriculum sat with education agencies and professional bodies captured by constructivist ideology. The consequences – illiterate school-leavers, foreclosed futures – fell on children, families, and employers who had no power to hold the system accountable. Minister Stanford’s reforms are the first serious attempt to reverse the decline. Yet they have met fierce resistance – from the very people who held authority but never bore consequences.

Health spending rises year after year. Outcomes do not follow. Patients waiting more than four months for treatment grew from 4,300 in January 2019 to over 37,000 by January 2025 – nearly ninefold in six years. Only around 61 per cent of patients now see a specialist within the four-month target, well short of the 95 per cent goal. The human cost is measurable in deteriorating conditions, delayed diagnoses, and preventable deaths.

District Health Boards – now consolidated into Health New Zealand – were funded for activities, not results. Hospital managers faced no penalty for rising wait times and no reward for reducing them. The metrics that mattered were financial: staying within budget, managing headcount, processing volume. A chief executive whose waiting lists doubled faced no career consequences. One whose budget blew out did.

In housing, affordability collapsed not through market failure but government architecture. Councils approve development, but growth costs – pipes, roads, stormwater, parks – fall within their budgets. Meanwhile, revenues from population growth – GST on construction, PAYE from new residents, company tax from expanding firms – flow to central government.

As Luke Malpass and Michael Bassett showed in Free to Build, this fiscal mismatch makes every new household look like a cost, not an opportunity. Saying no becomes the financially rational choice. Councils aren’t hostile to growth. They’re responding to an architecture that punishes them for enabling it.

In regulation, failure operates at two levels, and authority is separated from consequences at both. When Parliament makes new regulations, officials preparing regulatory impact statements face weak incentives to get analysis right. A poor RIS rarely stops a bill. Meeting deadlines and avoiding embarrassment are what matter most. The costs of badly designed regulation – higher prices, slower approvals, stifled innovation – fall on firms and consumers, not on officials, ministers, or Parliament. The results are visible in New Zealand’s decline on the OECD’s Product Market Regulation survey – from 2nd place in 1998 to 20th today.

When independent regulators exercise delegated powers, the accountability gap can be even wider. Bodies like the Commerce Commission wield extraordinary authority with limited internal checks and limited external scrutiny. As I documented in Who Guards the Guards? and Reassessing the Regulators, the consequences were predictable: inconsistency, unpredictability, and a striking lack of respect from those they regulate.

How the Architecture Shapes the Public Service

The domain-specific failures of education, health, housing, and regulation each have their own causes. But the same architectural logic – authority separated from consequences – shapes how the entire public service operates.

Ministerial authority is diluted by design. The Public Service Act 2020 placed the Public Service Commissioner between ministers and their departments. Chief executives now sit on a “public service leadership team” chaired by the Commissioner, who is also their employer and the person primarily responsible for their promotion. The Act’s language is revealing: its purpose includes enabling “both the current Government and successive governments” to implement policies – as if the public service serves some abstraction called “government” rather than the ministers voters actually elected. As I have previously argued, this drives a wedge between the leadership of government departments and the ministers to whom they are supposed to report. The result is a structural separation of democratic authority from bureaucratic accountability. Ministers arrive with mandates; the machinery answers to someone else.

The Public Service Amendment Bill 2025, currently before Parliament, takes some steps in the right direction. It clarifies that chief executives are responsible to their ministers, strengthens ministerial involvement in performance reviews, and removes language that prioritises collective spirit over individual accountability. But it leaves the fundamental architecture intact. The Commissioner still chairs the public service leadership team, sitting between departmental heads and their ministers.

Meanwhile, fragmentation disperses accountability. With 81 ministerial portfolios and 43 departments, New Zealand has three times as many portfolios and nearly twice as many departments as peer nations like Ireland or Singapore. Splitting housing across four portfolios doesn’t create better policy. It creates four escape routes. As Jemma Stevenson and I showed in our 2025 report for the New Zealand Initiative Unscrambling Government, when things go wrong, everyone can plausibly blame someone else. Since then, the government has announced a new ministry consolidating housing, transport, and environment functions, but the broader architecture of government remains hopelessly fragmented.

Expertise deficits compound the problem. The public service has elevated generalist management over subject-matter expertise. Officials rotate frequently between departments to demonstrate “leadership potential.” The Public Service Act 2020 formalised this by requiring “flexible deployment.” The result is administrators who know how to run processes but may lack the depth to challenge flawed assumptions or spot emerging failure. Career incentives reward mobility, not mastery.

Perhaps above all else, input measurement leaves outcomes invisible. The Budget measures inputs and activities, not whether literacy rises, waiting times fall, or houses become affordable. Better Public Services delivered real improvements under the Key government, but disappeared with the government that introduced it. Social Investment attempted to reorient spending toward long-term outcomes but was politically orphaned under Labour. The current coalition government has revived outcome targets across health, education, and other domains – a welcome step. The test is whether this approach becomes embedded in statute and convention rather than disappearing with the next change of government. Unless the habit of accountability takes hold, nothing will change.

As the Coster-McSkimming affair revealed, silence at the top can keep problems hidden. As I argued in a recent Herald column, the incentives for senior officials often point away from disclosure and toward institutional protection. When exposing a failure exposes your own earlier inaction, and failing to expose it carries no risk, silence becomes the safer choice.

In Demystifying the State, Tony Burton documented how this problem contributed to the scandal of child abuse in state care: “What is striking is how career-enhancing it was to protect departments at the expense of victims of state child abuse. State lawyers have become judges, department chief executives have received state honours.”

What Must Change

This architecture has accumulated through decades of decisions, each of which may have made some sense. But collectively, they have separated power from accountability. Reversing course requires reform on at least five fronts.

Restore ministerial authority: The Public Service Amendment Bill is a useful correction, but it still leaves coordination of the bureaucracy effectively mediated by the Public Service Commissioner through the leadership team. When cross-agency coordination is needed, it should be directed by elected ministers in Cabinet and enforced through chief executives’ performance agreements – not delegated to an unelected official. Clarity of command is clarity of accountability.

Defenders of the status quo will invoke the need for public service neutrality. But as I have argued elsewhere, senior appointments are already political – ministers shape appointment panels and Cabinet can reject candidates. The myth of neutrality simply obscures where accountability lies. Australia has acknowledged this reality, allowing incoming governments to appoint department heads aligned with their democratic mandate. New Zealand should do the same.

Restructure the executive: The ministerial maze must go. Eighty-one portfolios and 43 departments for a country of five million is absurd. Responsibility is spread so thinly that when things go wrong, everyone can plausibly blame someone else. Consolidating portfolios into coherent groupings, creating statutory junior minister roles to allow delegation without fragmenting the responsibility of senior ministers, and halving the number of departments would make accountability real. When failure has an address, behaviour changes.

Rebuild expertise: Career paths must reward subject-matter expertise, not favour managerial rotation. The Public Service Act 2020 elevated mobility over mastery and collective spirit over individual accountability. These may sound like virtues. In practice, they dissolve responsibility.

Measure what matters: This is the deepest change. Public sector reporting must shift from inputs to outcomes. Parliament cannot hold the Executive to account without meaningful information about what spending or regulation achieves. This is not a call for crude targets that distort behaviour. It is a call for honest measurement. Are more children reading at grade level? Are waiting times falling? Is housing becoming more affordable? These questions should be central to every agency report, every ministerial assessment, every Budget debate. Outcome measurement must be embedded in statute and convention, not left to the preferences of whoever holds office. Critics will warn that what gets measured gets gamed. They are right – but the answer is better measures, not measurement avoided.

Make suppression risky: Senior officials who bury serious allegations currently face no criminal sanction. A modern, narrowly drawn statutory offence – a New Zealand equivalent of misconduct in public office – would change the calculation. The threshold should be high and safeguards robust. But when those with power to surface the truth choose to suppress it, that choice should carry real risk.

Why all this matters

New Zealand cannot afford institutions that systematically produce failure. Yet that is what we have built. Not through malice. Not through incompetence. But through an architecture that makes failure rational and success heroic.

The economists who mapped this territory were not fatalists. They studied failure not to excuse it but to explain it – and explanation opens the door to reform. Systems that align authority with consequences work better than systems that separate them. That is evidence, not ideology.

New Zealand’s institutional failures are the predictable result of systems that reward the wrong things and punish the right ones. We have built a state that treats spending as success, process as performance, and silence as safety.

Change the incentives and behaviour will change – and outcomes along with it. Leave the architecture untouched, and nothing will.

If we want a state that works, we must build one whose incentives make success – not failure – the rational choice.