The Free and Frank Series: The Public Administration Toolkit Assembled So Far

20/12/2025

This is the eleventh instalment in a series exploring what free and frank advice means in Aotearoa’s constitutional arrangements. Over the previous nine posts, I have assembled a conceptual toolkit: the public service bargain, the purple zone, the authorising environment, and the political economy that underlies them all. This week, I pause to gather these pieces together and to offer a first signal of what the empirical research, involving 92 practitioners across the last nine administrations, has begun to reveal. By the end of this post, I hope to have shown how the toolkit fits together and why the structured disagreement it helped me find may tell us something fundamental about how democracy works when nobody is watching.

Four years ago, I asked what I thought was a simple question: what is free and frank advice? The phrase haunts our democracy, invoked in our statutes, our Cabinet Manual, our daily political discourse, yet it remains stubbornly, fascinatingly undefined. I thought I was looking for a definition. I was wrong. Spectacularly, fascinatingly wrong. What I found instead was a ghost in the machine: a concept whose power lies precisely in its ambiguity. This series has been the record of my journey to understand why. So far, we have assembled a conceptual toolkit to make sense of this mystery. Today, I want to show you how the pieces fit together and reveal something fundamental about how power works when no one is watching.

The Puzzle

Free and frank advice occupies a unique position in Aotearoa’s constitutional arrangements. Unlike other Westminster jurisdictions, we have never had a major constitutional crisis that forced us to define precisely what we mean by the relationship between ministers and officials. We have improvised our way through constitutional development, creating conventions and understandings that work in practice but resist precise definition.

The research I have been conducting, involving 92 practitioners across the last nine administrations, elected and unelected, using interviews, Q-methodology, workshops and wānanga, to surface the systematic patterns in how people understand what free and frank advice is and why it matters, reveals something striking. The disagreement about what free and frank advice means is not random. It is structured. Six distinct philosophical positions emerge from the data, each representing a coherent way of understanding how advisory relationships should operate. Like tribes speaking the same language but meaning entirely different things. But I am getting ahead of myself: more on that in March. Suffice it to say, this is not confusion. It is a constitutional design, whether conscious or unconscious.

The Toolkit We Forged

Our journey began with the public sector bargain: the invisible, unwritten contract that governs the relationship between ministers and officials. Christopher Hood’s work revealed that these arrangements operate through implicit deals: officials surrender certain political rights and public visibility; in return, they receive permanence, professional autonomy, and protection through ministerial responsibility. The classical bargain gave officials the security to speak uncomfortable truths, knowing they would still be present when the minister had moved on.

But that bargain fractured. Donald Savoie’s Breaking the Bargain documented the collapse across Westminster systems. Fixed-term contracts replaced permanence. Performance agreements replaced professional trust. The result was officials caught in an impossible position: expected to be simultaneously independent and subordinate, both frank and compliant, responsive yet impartial. Dennis Grube’s research showed officials increasingly drawn into public visibility to defend government positions: a fundamental breach of the anonymity that once made frankness professionally safe.

In Aotearoa, the fracturing was particularly intense. The State Sector Act 1988 replaced permanent tenure with fixed-term contracts for chief executives, introducing what Richard Norman aptly termed the spectre of “obedient servants”: officials who might minimise frankness when their career safety depends on ministerial satisfaction.

The bargain sets the stage, but it must be operationalised somewhere. That somewhere is the purple zone: the institutional marae-ātea where the blue of politics and the red of administration bleed into one another. Alex Matheson and Howard Fancy named this space in 1995, recognising that the traditional dichotomy between political and administrative roles never worked cleanly in practice. The purple zone is where strategic conversation creates shared meaning, where political will becomes administrative action.

But the purple zone is not static. It is furnished differently by successive governments, and that furniture shapes what advice must contain to be considered adequate.

Under Bolger and Shipley, Strategic Result Areas crossed portfolio boundaries with explicit targets for reducing Māori disparities—embedding Article Three Te Tiriti obligations into the purple zone’s integrative machinery. For officials offering advice on any significant initiative, this infrastructure made certain questions mandatory. Advice that failed to address whether a proposal would improve or worsen Māori–non-Māori disparities was not merely incomplete; it was inadequate. The purple zone furniture gave Te Tiriti procedural voice.

When Clark’s government replaced this framework with ground-breaking lifecycle approaches and opportunity-for-all framing, that procedural voice disappeared. The constitutional obligation remained, but the infrastructure that made honouring it procedurally expected rather than personally risky was removed. Under Key and English, elaborate performance frameworks and Better Public Services results privileged evidence-based professional authority: officials had systematic channels for frank advice about what was working. Under Ardern, dramatic simplification around wellbeing and child poverty left political will unmoored from the administrative capability infrastructure that might have supported it. These are simplifications, of course, but are offered as illustrations.

The picture they paint is of the essential quality of free and frank advice that may remain constant across these configurations. But the form it takes, the questions it must answer, the framing it must use to engage ministers, all shift as governments reconfigure the workspace.

Rose Cole’s doctoral research locates this purple zone precisely: the minister’s office, staffed by a mix of partisan and non-partisan advisers. Portfolio private secretaries, public servants seconded from departments, must serve dual principals simultaneously: the minister as immediate authority, and the department as employer and career home. They navigate what Cole’s participants called “a very dicey minefield every single day,” developing “the real art” of not burning bridges whilst still getting stuff done. This is not dysfunction. This is the daily lived reality of working in constitutional ambiguity.

But who gets to decide what a good bargain looks like, or how the purple zone should be configured? That question leads us to the authorising environment. It pushes beyond the simple minister-official relationship to ask whose legitimacy counts. Officials must navigate not two but many sources of authority: ministers, departments, Parliament, future governments, Te Tiriti obligations with Iwi and Hapū, professional standards, statutory requirements, and the broader public interest. Each creates different expectations. What satisfies ministerial authority might undermine professional standards. What honours Te Tiriti obligations might conflict with immediate political priorities.

Mark Moore’s concept of “public value” offered one answer: officials as wise stewards discerning shared public interest through consultation and engagement. But as Rhodes and Wanna warned, this risks giving officials interpretive authority over what counts as valuable, blurring the line between bureaucratic judgement and democratic mandate. The question of which authority possesses legitimacy when they fundamentally conflict is not one officials can resolve; it is a ministerial prerogative. Perhaps free and frank advice remains underdefined so that officials can navigate between competing authorities whilst maintaining enough shared meaning to enable coordination, accountability, and endurance.

Finally, we arrived at political economy: the concept that sits beneath all the others. If the purple zone is the workspace, political economy explains how that workspace gets furnished, whose interests the furniture serves, and why some arrangements persist while alternatives fade from view.

Three traditions escalate the analysis. The market-institutional tradition (Polanyi, Barma, Vogel) shows that markets are built, not natural: “laissez-faire was planned.” The critical tradition (Kelsey, Roper) asks the harder question: built by whom, and for whose benefit? The Māori political economy tradition (Scobie et al.) reveals something more fundamental still: that alternative economic logics operate according to entirely different rules, organised around taonga, mahi, and utu, where wealth is measured not by what accumulates but by what passes through in service of others. This raises questions not merely about who controls construction, but whether a single institutional framework can accommodate genuinely plural modes of economic organisation.

Political economy, potentially, shapes what can be said, what goes without saying, and what cannot be said without threatening the arrangements that hold a settlement together. Some advice is structurally unsayable—not because it is unwelcome but because it would challenge the credibility arrangements through which political coalitions maintain themselves.

The Machine Exposed

The four concepts above are not a simple list; they are a nested and layered analytical framework. Political economy provides the deep structure, furnishing the purple zone with its assumptions, its biases, and its silences. The bargains struck between ministers and officials are negotiated within that furnished space, not on a level playing field. And the entire arrangement is held accountable, or fails to be, to the multiple, competing sources of legitimacy in the authorising environment.

This is why a single, static definition of free and frank advice may be impossible. What counts as “frank” advice is determined by the furniture in the room. Advice that challenges the political economy settlement itself, the very way the room is furnished, is not merely unwelcome; it is structurally difficult to voice.

A Statutory Whakapapa

The legislative trajectory reinforces this analysis. Parliament has progressively elevated free and frank advice from unspoken convention to constitutional principle, whilst never once defining what “free,” “frank,” or their conjunction actually means.

The Official Information Act 1982 made “free and frank expression of opinions” a withholding ground, treating the concept as self-evident, something officials would intuitively recognise. The State Sector Act 1988 shifted it from protected information to a contractual deliverable, requiring chief executives to maintain the capacity for such advice. The 2013 amendments specified the “tendering” of free and frank advice, still treating it as a product, an output to be quality-assured.

The Public Service Act 2020 made a subtle but significant shift. Rather than requiring the tendering of advice, it established that a core principle of the public service is to act “in a free and frank manner” when giving advice. Note the word: manner. Parliament moved from specifying what advice should contain to prescribing how officials should behave when providing it. From noun to verb. From product to person, and in particular, to the way the person behaves in a specific moment.

Whether this represents progress depends on what you think the problem actually is. If officials do not know what freeness and frankness require, prescribing behaviour offers little help. If they know perfectly well but find it professionally hazardous to deliver, prescribing behaviour without changing the conditions that make frankness risky amounts to exhortation without protection.

Why Ambiguity May Be Constitutive

Recent work by Richardson, Durose, Cairney, and Boswell on policy “buzzwords” offers a framework for understanding why free and frank advice differs from concepts like prevention or co-production that cycle through fashionable obscurity. Buzzwords catch fire partly because their ambiguity generates broad appeal, but that same breadth makes them challenging to operationalise, leading to eventual abandonment.

Free and frank advice shares these characteristics but has proven remarkably persistent, adapting rather than fading. Perhaps this is because it operates in two fundamentally different registers simultaneously: as confidential counsel to ministers (requiring trust and discretion) and as a contribution to public policy debate (requiring openness and contestability). Richard Mulgan drew this distinction carefully: analysis and advice are different. Policy analysis can be published, peer-reviewed, and publicly contested, whereas policy advice requires confidentiality precisely so that serious disagreement between ministers and officials can remain hidden.

Policy advice, in Mulgan’s view, is a structure for debate: the starting point for a decision. Thus, perhaps, the moment you define free and frank advice too precisely, you force a choice between these registers. Any definition privileges certain authorities whilst subordinating others.

The Wisdom in Ambiguity

So, for now, we are left not with a definition, but with a sharper understanding of the problem. The ambiguity of free and frank advice may not be a bug. It may be a feature. It may be what allows the advisory relationship to adapt to shifting political economies and evolving purple zones without shattering: a constitutional concept that preserves enough shared meaning to enable coordination and accountability whilst accommodating the six distinct philosophical positions my research has surfaced.

But this constitutional wisdom is purchased at a cost. Individual officials, often junior, often in precarious positions, bear the daily weight of navigating ambiguity without procedural support. The “real art” that Cole’s participants describe is not taught in induction programmes or specified in guidance documents. It is learned through observation, through accumulated experience of watching which advice proves influential and which disappears without trace.

The Journey Ahead

The toolkit is now assembled. In the coming weeks, we turn to the political theory that underpins these arrangements. We will examine four traditions: consent theory, rationalist deliberation, non-domination and self-government, and trust-based accounts: each of which deepens rather than solves the definitional problem. We will arrive at agonistic democracy and Bonnie Honig’s concept of “remainders”: the idea that democratic life is constituted through contestation, and that some conflicts cannot – and should not – be finally resolved without doing violence to the democratic project itself.

From there, we turn to methodology: to Bevir and Rhodes’s interpretive approach, and why understanding how the 98 practitioners I interviewed construct meaning around contested concepts requires attending to beliefs, traditions, and dilemmas rather than simply mapping institutional structures.

And then, finally, to the empirical findings themselves. The six factors. The tribes. What they reveal about how free and frank advice actually operates in practice, and why the disagreement itself may be the answer.

The ghost remains a ghost. But we can now see the shape of the machine it haunts. We now turn to the political theory.