The Free and Frank Series: He Whakapapa

06/12/2025

This is the ninth instalment in a series exploring what free and frank advice means in Aotearoa’s constitutional and institutional arrangements. Having considered what the census data reveals about practitioner engagement with the concept, this week I trace the statutory whakapapa of the concept: its journey from unspoken convention before 1982, to protected information under the Official Information Act, to contractual output under the State Sector Act 1988, to behavioural principle under the Public Service Act 2020. By the end of this post, I hope to have shown that, over four decades and across five pieces of legislation, Parliament has progressively elevated free and frank advice to a constitutional principle whilst maintaining deliberate silence on its substance. That silence, I would suggest, is not an oversight.

Today’s post follows the statutory journey from the Public Service Act 1912, which made no mention of advisory duties, through the Official Information Act 1982, the State Sector Act 1988, the 2013 amendments, and the Public Service Act 2020. At each stage, Parliament elevated the importance of free and frank advice whilst maintaining silence on its substance, culminating in a shift from noun to verb: from specifying what advice should contain to prescribing how officials should behave when providing it.

If you have worked in the public service for any length of time, you will have encountered the phrase “free and frank advice” in induction materials, guidance documents, and earnest kōrero over coffee. You may have nodded along, assuming everyone else understood what it meant. You may have wondered, privately, whether you were the only one who found the concept slippery: clear enough in the abstract, but maddeningly elusive when you needed to know whether this particular piece of advice, to this particular minister, in this particular moment, qualified as free and frank advice. You are not alone. Parliament has been grappling with the same problem for decades. What follows is a trace of the concept’s statutory whakapapa: how “free and frank advice” moved from unspoken convention to protected information to contractual deliverable to behavioural principle, without anyone, at any point, specifying what the words actually mean.

There is something peculiar about free and frank advice in Aotearoa New Zealand. Over more than a century, Parliament has progressively elevated this concept from unspoken convention to constitutional principle, while never once defining what “free,” “frank,” or their conjunction actually means. This legislative silence is worth sitting with. It is not, I suspect, accidental.

Before the Words Appeared

The Public Service Act 1912 and the State Services Act 1962 made no mention of any duty to provide policy counsel to ministers. The statutes constructed a public service premised on administrative neutrality and operational efficiency¹. Advice, if it appeared at all in the legislative imagination, was simply what officials did.

This absence did not mean advice was not given: quite the contrary. It reflected a constitutional understanding where the advisory function was governed by unwritten Westminster conventions, emerging from practice rather than imposed through formal declaration.

Officials learned what constituted appropriate advice through observation and accumulated experience. They watched how advice was received, which topics were sensitive, and what forms of challenge were tolerated. This tacit knowledge could not be codified precisely because its value lay in contextual application: knowing when to push back and when to defer, how to signal disagreement whilst maintaining a functional working relationship.

The concept existed. It simply did not need to be named.

1982: Protection Through Obscurity

One of the first statutory appearances of “free and frank” came not as an obligation but as something requiring protection from disclosure. The Official Information Act 1982 established the “free and frank expression of opinions” as a withholding ground, presumably to shield ministerial deliberations from public scrutiny.

The underlying assumption was revealing: officials would intuitively recognise what constituted “free and frank” advice and, on that basis, know when to withhold that particular bit of information. Parliament treated the concept as self-evident, requiring no elaboration.

Yet this legislative move carried consequences that took years to become visible. By making documented advice potentially subject to public release, the Act began shifting counsel from written to verbal form. Within a decade, officials were already crafting briefs with an eye to how they might appear if disclosed: developing what would become the “front page of the Dominion test.” Nicola White’s scholarship on this is invaluable, to say the least (2007).

The convention of free and frank advice did not disappear. It migrated into spaces where documentation could be avoided, fundamentally altering the evidentiary record of policy deliberation whilst leaving the formal commitment to candour ostensibly intact.

1988: From Convention to Contract

The sweeping reforms of the 1980s fundamentally reconceptualised the relationship between ministers and officials. The State Sector Act 1988 made chief executives employers in their own right, appointed on fixed-term performance contracts. It shifted free and frank advice from protected information to direct responsibility: first requiring chief executives to maintain organisational capacity for such advice, and later mandating that it be tendered to ministers.

This seemingly technical clarification represented something more profound. By treating policy advice as a purchasable output to be specified by contract and measured by ministerial satisfaction, it introduced the transactional logic of the market into the core of the advisory relationship.

The irony was acute. A framework designed to enhance accountability created powerful incentives for defensive caution. When career progression depends on ministerial satisfaction ratings, officials need not be instructed to align analysis with political preferences: the feedback loops embedded in assessment processes teach this lesson through accumulated experience of watching which advice proves influential and which disappears without a trace.

2013: Legislating Quality

By the early 2010s, persistent concerns about declining advice quality prompted direct legislative intervention. The State Sector Amendment Act 2013 specified the character of the deliverable itself, officially requiring the “tendering of free and frank advice.”

Yet this approach contained a conceptual flaw. By specifying “free and frank” as a characteristic of the deliverable, it continued to treat advice as a product: something that could be defined, measured, and quality-assured like any other output.

The amendment, while useful, failed to recognise that candid counsel is not an output to be specified but an outcome of relational and cultural conditions that make up our constitutional culture: trust, psychological safety, a shared commitment to the public good, and pragmatism; the aggregate of the behaviours within which the advisory relationship and the policy advisory systems operate.

2020: From Product to Behaviour/Process

The Public Service Act 2020 represents the most recent legislative iteration and makes a subtle but significant shift. Rather than requiring the “tendering” of advice and that transactional language of information products, 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. Not ethos, not disposition, not character, but manner. The dictionary is unambiguous here. Manner means “a way in which a thing is done or happens.” It describes behaviour, conduct, a process: an observable way of doing something.

This matters. Parliament moved from specifying what advice should contain to prescribing how officials should behave when providing it. The focus shifted from the product to the person: from the noun to the verb.

Whether this represents progress depends on what you think the problem actually is. If the difficulty lies in officials not knowing what free and frank advice contains, then prescribing behaviour offers little help.

For if the difficulty lies in officials knowing perfectly well what freeness and frankness require but finding it professionally hazardous to deliver, then prescribing behaviour without changing the conditions that make frankness risky amounts to exhortation without protection.

What the Trajectory Reveals

This legislative journey exposes a persistent tension. Each statutory iteration increased the functional importance of free and frank advice whilst maintaining silence on its conceptual form. Parliament has elaborated that such advice matters without ever specifying what it actually is.

Perhaps this is not legislative oversight but constitutional wisdom. Any rigid definition, stipulating either that officials must always defer or that ministers must always accept professional advice, would privilege one side over the other, foreclosing legitimate contestation rather than managing it.

It is possible that the concept’s enduring ambiguity creates space for ongoing, context-specific negotiation over where the appropriate balance lies. But it also leaves practitioners navigating competing interpretations without authoritative guidance, and the path of least resistance: a responsive service that ministers welcome, which carries professional reward whilst its alternative carries risk.

The statute has told us that manner matters. It has not told us what manner to adopt when the minister does not wish to hear what we have to say: the behaviour of someone being free and frank.

The statutory trajectory exposes a persistent pattern. Each legislative iteration increased the functional importance of free and frank advice whilst maintaining silence on its conceptual form. Parliament has told us that manner matters. It has not told us what manner to adopt when the minister does not wish to hear what we have to say. The legislative silence may not be oversight. It may be the space in which the concept does its democratic work.

References

White, N. (2007). Free and frank: Making the Official Information Act 1982 work better. Institute for Government and Policy Studies, Victoria University of Wellington.

¹ The usual discourse positions efficiency as an import of the 1980s reforms, as though New Public Management invented the concept and imposed it upon a public service previously unconcerned with such matters. The archival record tells a different story. If you spend any time with the reviews and reports of the 1940s and 1950s, the word “efficiency” appears with striking regularity. The Public Service Commission of that era was deeply preoccupied with value for money, with eliminating waste, and with ensuring that the machinery of government operated without unnecessary friction. Our scholarly obsession with NPM and its manifold problems has created a kind of historical amnesia. We treat the 1980s as a rupture, a moment when market logic colonised a public service that had previously operated according to entirely different values. The reality is messier. Ideas of efficiency, effectiveness, and value for money were long-standing obsessions in old public management before they became orthodoxies of new public management. What changed in the 1980s was not the presence of these concerns but their theorisation, their elevation into a comprehensive reform programme, and their embedding in contractual and performance management architecture. The words were already there. The machinery built around them was new.