Free and Frank Advice: Beyond Aotearoa, how other systems carry candour

11/10/2025

I compare how Australia, the United Kingdom, Canada and Aotearoa carry the obligation of candid advice, and I argue that beneath four different arrangements the four systems share one silence about what that obligation requires.

This is the second of twenty posts drawn from my doctoral thesis on free and frank advice in Aotearoa. In this post I step outside Aotearoa to look at how three other Westminster systems carry a kindred obligation of candour, and at what the Australian experience, Robodebt in particular, shows about a convention left undefined.

Free and frank advice, or a recognisable equivalent of it, appears well beyond Aotearoa, and across the Westminster family there is a kindred expectation that a permanent, merit-based and apolitical public service owes the Government of the day its best professional judgement, even when that judgement is unwelcome.

In this post I do three things: first, I look at Australia, which codified the convention and altered it in the codifying; second, I look at the United Kingdom, which declines to name the convention, and at Canada, which sets candour and loyalty apart beneath a democratic value; and third, I set out what the four systems share, and what the Australian mirror shows about the cost of that shared silence.

Australia: codification and its repair

Australia is the closest sibling of Aotearoa, and its original formulation was “frank and fearless”, a phrase that shares the alliterative emphasis of its trans-Tasman counterpart and that for the better part of a century governed the Australian service as an unwritten convention among the traditional Westminster values (MacDermott, 2008; Mulgan, 2012). Australia then moved beyond convention far earlier than Aotearoa did, codifying its public service values in the Public Service Act 1999 (Cth), and in my reading the act of codification changed the convention it set out to capture.

The original Public Service Bill 1997 asked only that the service give the Government timely advice, a bare clause stripped of the qualifiers that had been recommended, and fearlessness, as MacDermott puts it, “was not even on the agenda” (MacDermott, 2008, p. 4). What emerged in the Act was a single value under which the service was responsive to the Government in providing frank, honest, comprehensive, accurate and timely advice and in implementing the Government’s policies and programs (Public Service Act 1999 (Cth), s. 10(1)(f), as enacted). By gathering the giving of advice and the carrying out of policy under responsiveness, the statute asked the official to be frank and responsive at once, and it left the moment the two duties part company for the official to work out, case by case.

Australia later undid that fusion, and the way it did so is instructive. The Public Service Amendment Act 2013 (Cth) replaced the values in full and placed frankness under impartiality, so that the service is apolitical and gives the Government advice that is frank, honest, timely and based on the best available evidence (Public Service Act 1999 (Cth), s. 10(5)). Implementation moved elsewhere in the values, and an epistemic standard of evidence replaced the older demand for comprehensive and accurate advice. In 1999 the official was frank in order to serve, and after 2013 the official is frank because the service stands apart from politics. What the redrafting kept, in my view, was the silence, because the value still asserts that advice will be frank while saying nothing about what frank advice is, and the over-responsive culture that had formed under the fused value was still in place when the Robodebt scheme ran its course under the repaired wording.

The United Kingdom and Canada

The United Kingdom reaches the same obligation by a different path, and the words “free and frank” appear in neither its statute nor its Code. The Constitutional Reform and Governance Act 2010 and the Civil Service Code instead set out four core values, integrity, honesty, objectivity and impartiality, and the Code gives objectivity a working content that sits close to what Aotearoa and Australia call frankness: officials must give advice on the basis of the evidence and present the options and facts accurately, must not ignore inconvenient facts, and must not frustrate the implementation of policies once decisions are taken (Civil Service, 2015; Constitutional Reform and Governance Act 2010, s. 7). A code of that kind can name the duty and name its limit, and in my reading it still leaves the official to judge where the one becomes the other, with no rule to reach for.

Impartiality completes the construction, and the Code splits it into fairness between persons and interests on the one hand, and political impartiality on the other, which asks officials to serve whatever government is returned and to keep the confidence of ministers while keeping their capacity to serve a future government (Civil Service, 2015). The nearest the United Kingdom comes to free and frank advice, as I read it, is through the discipline of objective evidence and the even hand of impartiality, while the Australasian path runs through frankness and the courage to use it.

Canada sets the obligation on its own terms again, and its Values and Ethics Code keeps the loyal carrying out of lawful decisions and the giving of open, candid and impartial advice as separate expected behaviours, gathered beneath a single value, Respect for Democracy (Treasury Board of Canada Secretariat, 2012). Loyalty and candour sit one behaviour apart and both answer to democracy itself. Canada is also alone in the family in naming the minister’s side of the bargain, recording in the Code’s preliminary text that ministers play a critical role in supporting the responsibility of public servants to give professional and frank advice. The system that performs these behaviours is the most centralised of the four, with a powerful prime minister’s office and the largest complement of political advisers (Craft & Halligan, 2020). In my view the Canadian Code states an ideal its own executive strains against, and nothing in the Canadian design failed, because what changed sat outside the Code while the text of the obligation stayed where it was.

One silence across four settings

Aotearoa names free and frank advice in its own right, as one of five public service principles in the Public Service Act 2020, each a means to the purpose the Act lays down, and that placement marks it out from the other three (Public Service Act 2020, s. 12(1)(b)). Yet for all that explicit naming, the statute leaves unsaid what free and frank advice is, and the comparison shows the same silence running through all four systems: each carries the obligation in a different institutional form, and none says what free and frank advice, or frankness, or candour requires of officials and ministers (McLean, 2024). I think a family of systems relying so heavily on a convention all of them have left unspecified is best treated as a finding in its own right, and the four settings point to one underlying condition, a formal commitment to free and frank advice running ahead of the conditions under which it can safely be given.

Australia holds the warning mirror, because it is the near case, sharing the constitutional inheritance and reforms that ran along similar lines, and its divergence turns on a small number of identifiable choices. The New Public Management reforms ran harder there, and six departmental secretaries lost their posts on the change of government in 1996, while Aotearoa has kept chief executive appointments at arm’s length from politicians (Dowding & Taflaga, 2025; MacDermott, 2008). Responsiveness is proper up to a point, and it tips into excess when officials trim or redraft their own recommendations to match the minister’s, lending a false appearance of independent advice (Mulgan, 2008). With fearlessness left to convention and responsiveness still powerful in the administrative culture, the mirror shows what becomes of candour when the courage to use it goes unprotected.

Robodebt is where that over-responsiveness turned catastrophic, and the Royal Commission set down the record of a convention that had stopped working, with officials who had concerns staying silent or going unheard, legal advice put aside, and compliance with ministerial preference standing in for candid assessment, and one reading of that record concludes that “departmental advice tended to be neither frank nor fearless” (Mulgan, 2024, p. 5; Royal Commission into the Robodebt Scheme, 2023). Other readings treat the failure as deliberate, the product of an overly responsive culture, and trace it to secretaries stripped of tenure and brought under firmer prime-ministerial control (Henman, 2025; Podger, 2023). The cost fell on welfare recipients pursued for debts wrongly raised against them, some of whom died, and in my view that is the plainest reminder that when an advisory system fails to surface what decision-makers must hear, those with the least power pay.

The Australian literature also presses into the terms of the convention itself, treating the phrase as useful shorthand that obscures the reasons officials may disagree and the degrees of push-back different situations call for (Mulgan, 2024). In my reading frankness and fearlessness answer to different parties, because frankness belongs to the work of service, telling ministers what they need to hear, while fearlessness is a check on responsiveness, exercised where what the minister wants and what the public interest requires have come apart, and its point is that the minister will resist it. Joining the two words joins a duty owed to the minister with a limit imposed for someone else’s benefit, and I read the phrase as making it harder to see where the one ends and the other begins.

Aotearoa is increasingly treated as exemplary in this family, judged best placed among Westminster systems to withstand pressure on impartial frank advice and commended for the merit-based, arm’s-length appointment of its public service leaders under the 2020 Act (Aucoin, 2012; Dowding & Taflaga, 2025). To be treated as exemplary in a convention that still lacks a practitioner account seems to me precisely the reason to understand it. The comparison establishes two things: free and frank advice is carried across the Westminster family, so the puzzle is no local accident of drafting, and no member of the family has said what the obligation requires, so no better-drafted statute elsewhere has already solved it.

In the next post I turn to the inheritance itself, the Westminster frame, the Whitehall service and the unresolved relationship with Te Tiriti o Waitangi through which free and frank advice reached Aotearoa.

References

Aucoin, P. (2012). New political governance in Westminster systems: Impartial public administration and management performance at risk. Governance, 25(2), 177–199.

Civil Service. (2015). The Civil Service Code. GOV.UK.

Constitutional Reform and Governance Act 2010 (UK), c. 25.

Craft, J., & Halligan, J. (2020). Advising governments in the Westminster tradition: Policy advisory systems in Australia, Canada, New Zealand and the United Kingdom. Cambridge University Press.

Dowding, K., & Taflaga, M. (2025). Bad governance in Australia and how to mitigate it. Australian Journal of Public Administration, 84, 539–557.

Henman, P. (2025). Robodebt, cultures and useful idiots: Why Robodebt was not a techno-failure. Australian Journal of Social Issues. Advance online publication.

MacDermott, K. (2008). Whatever happened to frank and fearless? The impact of new public management on the Australian Public Service. ANU E Press.

McLean, J. (2024). Conventional norms and the comparative constitutional law project. Comparative Constitutional Studies, 2(2), 272–291.

Mulgan, R. (2008). How much responsiveness is too much or too little? Australian Journal of Public Administration, 67(3), 345–356.

Mulgan, R. (2012, May 30). What future for free and frank advice? [Address]. Institute of Public Administration New Zealand Centennial Lecture, Wellington, New Zealand.

Mulgan, R. (2024). Ten lessons for public servants from Robodebt. Australia and New Zealand School of Government.

Podger, A. (2023). Report to the Royal Commission into the Robodebt Scheme.

Public Service Act 1999 (Cth).

Public Service Act 2020 (NZ).

Public Service Amendment Act 2013 (Cth).

Royal Commission into the Robodebt Scheme. (2023). Report: Volume 1. Commonwealth of Australia.

Treasury Board of Canada Secretariat. (2012). Values and ethics code for the public sector. Government of Canada.