Everybody says they want policy contestability because it tests ideas and can bring new insights and rigour to policy and decision making.

But there’s one thing my years conducting contestability on Australia’s defence and strategic policy from positions inside and outside government have taught me: it’s that Public Service mandarins only tolerate contestability if it doesn’t question their cherished assumptions and policy positions. And they certainly don’t want to establish a rigorous system of contestability that has the potential to fundamentally refute those assumptions and positions.

Again and again we see reviews that extoll the value of contestability, generally after a deplorable policy failure, and make recommendations that will supposedly bake effective contestability into the Australian government’s policy process. Then the mandarins make sure that when it comes to their pet policies and projects, effective contestability is not applied. Policy failure ensues, followed by reviews that say if only we had better contestability things wouldn’t go wrong and here’s how to do it. Rinse and repeat.

Let’s take the latest effort in this futile circle, the ‘Independent Review of Commonwealth funding for strategic policy work’ that the Government quietly released along with its response just before Canberra’s Christmas stand down. It’s better known as the Varghese review after its author, the former Secretary of the Department of Foreign Affairs and Trade, Peter Varghese, or to some in the media as the ‘get ASPI’ (i.e., the Australian Strategic Policy Institute) review.

It’s not entirely clear what prompted a current Public Service mandarin – the Secretary of the Department of the Prime Minister and Cabinet, Glynn Davis– to commission a former public service mandarin to conduct a review of public funding of external organisations by the national security community. There’s no particular trainwreck cited in the terms of reference and the sums of public money involved are trivial compared to many areas of Commonwealth expenditure.

But the review itself suggests the whole point is about ensuring contestability in policy making. Varghese starts with the mandarins’ usual homage to contestability stating that ‘the real value of contestability is that it can improve policy by explaining where a policy is not working or where its analytical foundations are either weak or wrong. Good policy making rests on a contest of ideas and the testing of assumptions.’

Varghese notes (quite correctly) that the contest of ideas isn’t happening, concluding that ‘Australian policy making tends to allow only a limited space for contestability from external sources. In theory, Governments favour policy contestability but for the most part where this occurs it is an internal exercise and, to the extent that external contestability occurs, it tends to be after policy has been made. This undermines the sector’s ability to provide alternative analysis or identify gaps in policy thinking, limiting it to an advocacy, educational, or critiquing role.’

Putting aside the key problem that there is very little space for internal contestability within government agencies either, this framing indicates that the solution is to open up space for ‘contestability from external sources, in particular before the Government has committed to a policy.

Unfortunately there is nothing Varghese’s recommendations or the Government’s response to it that will improve this situation, indeed quite few recommendations would further limit external contestability. It is yet another review produced under the Albanese Government that cements business as usual without even trying to challenge the status quo.

The fundamental irony in this review is that a key purpose of contestability is to expose actors’ cognitive biases, both conscious and unconscious, yet the review is riddled with Varghese’s unconscious biases as a former mandarin.

I’m not going to enter into the debate about whether this was a ‘get ASPI review’ or whether it implicitly or explicitly addresses the Chinese Communist Party’s 2020 list of 14 grievances against Australia, which included a call to shut down think tanks critical of China, but I’ll use one ASPI-related issue in the review to highlight the cognitive bias at work.

In the entire review only one individual is called out for criticism, the previous Executive Director of ASPI, who is not named but is easily identifiable as Peter Jennings, my former boss at ASPI and current colleague at Strategic Analysis Australia. Jennings’ sin was, to use Varghese’s term, ‘op-ed overreach’ which was based ‘on personal opinions rather than deep ASPI research’. That’s an odd claim considering Varghese himself praises the deep research that ASPI conducted under Jennings that informed his commentary. Nevertheless, Jennings’ public commentary, written for forums that people might actually read, was ‘partisan’ and ‘a failure’.

Now I disagree with Peter Jennings on many things; that’s the fundamental bedrock of contestability – the contestability that Varghese himself says we need and don’t have enough of. But here’s the mandarin’s bias at work: If opinion comes from within the system controlled by the mandarins, then it’s “informed” and part of the rigorous cut-and-thrust of ideas that, it is claimed, goes into good policy formulation. If it comes from outside (and worst of all, is aimed at the mandarins’ sacred cows such as deepening Australia’s economic reliance on exporting raw materials to China) then it’s deplorable ‘op-ed overreach’.

I worked for Varghese at the Office of National Assessment many years ago. Despite the firehose of intelligence feeds ONA received, that fact is that much of the assessment it provided to senior ministers was largely opinion. There’s no particular intelligence that can tell you what one country’s economic trajectory will be, or what the response of another country’s leadership will be to an event or potential event in our region.

ONA’s assessments were the views of public servants who had middling to excellent understanding of their subject matter, informed by their time working on that desk and their understanding of their audience’s interests. A bit like the opinion pieces in the better traditional media like The Economist or The New York Times. Or the writings of a former chief of staff to a federal minister, senior adviser to Prime Minister Howard and who also served as Deputy Secretary Strategic Policy in the Department of Defence—like Jennings, who was acting in accordance with the key tenet of Varghese’s review that ‘the real value of contestability is that it can improve policy by explaining where a policy is not working or where its analytical foundations are either weak or wrong.’

In fact, it’s a little hard to know what Varghese thinks good contestability should look like. Despite his condemnation of Jennings’ commentary lacking a basis in research, Varghese presents no evidence to support his own views about the value of different kinds of strategic policy work. Other than some listed interviews with people in the community—all establishment figures like Varghese—he presents no research to show what kinds of work have shaped policy, or improved ‘policy by explaining where a policy is not working or where its analytical foundations are either weak or wrong’.

At least Varghese, the Chancellor of the University of Queensland, admits that ‘academic papers are often too dense to inform policy development and lack forward leaning recommendations.’ But while he claims that these ‘deep research pieces tend to have a longer shelf life than the shorter policy-relevant pieces’, it is not reassuring that his view of the value of his institution’s work is the length of time its products spend gathering dust in the university library stacks.

Perhaps Jennings’ worst sin was that, regardless of whether you agree with his work, people actually read it and it had an impact beyond dense academic studies, or indeed the advice of the Canberra policy establishment. Indeed, one could argue that both the Coalition and Labor governments’ support for AUKUS shows that Jennings’ repeated early warnings about the Chinese Communist Party’s intent to upend the regional security order, had a significant impact.

Overall it doesn’t appear that Varghese has used his review to contest his own unconscious biases. The biggest problem with his report is that its “business-as-usual” recommendations won’t bring about the outcome it claims to seek, namely greater contestability in strategic policy making.

Australia’s national security mandarins have been a one-trick pony in recent decades; the solution to all our ills is greater co-ordination with the mandarins themselves as the coordinating body in the form of the Secretaries’ Committee on National Security (SCNS). It doesn’t matter what the challenge or threat is, greater oversight by SCNS will fix it. The fact that the policies SCNS has recommended and is implementing are often fundamentally weak (as sinners such as Jennings have pointed out) is irrelevant.

So it’s no surprise the review’s recommended solution to the problem of the lack of external contestability is to apply even more paternalistic guidance from Canberra’s mandarins meeting in their SCNS committee over our few security think tanks. I won’t recount the “paint-by-numbers” list of coordinating mechanisms that the review recommends here. We’ve seen this all before: benign words about “best practice governance” mask the reality of more bureaucratic oversight and funding control.

We’ve also seen what happens to contestability of any kind when it’s left to the mandarins to manage. We can look at a case study, admittedly of internal contestability, but it illustrates the key point that the mandarins proclaim their love of contestability until they don’t want it.

The previous government commissioned the First Principles Review (FPR) of Defence in 2014. This was meant to be the review to end all reviews—a definitive attempt, once and for all, to improve Defence higher management. Importantly the FPR emphasised the need for contestability. Indeed, ‘timely, contestable advice using internal and external expertise to provide the best advice so that the outcome is delivered in the most cost-effective and efficient manner’ was one of its foundational first principles. It made 47 references to contestability and three of its recommendations specifically aimed at institutionalising contestability in the Department. The Secretary and Chief of the Defence Force of the day certified to the Government that they endorsed and had implemented the FPR’s recommendation.

Remember: this was internal contestability, the kind that Varghese claims that our public service is better at. Yet the review failed virtually at the first hurdle. When the Department of Defence conducted the competition for the future submarine (eventually won by the French contender), the Department’s hierarchy excluded its internal contestability function from the process. This was Defence’s (and probably the Commonwealth’s) largest capital investment ever, yet there was no place for contestability in the process. Even though I managed the area of contestability that examined maritime projects at the time, I was never told why we were excluded, other than hearing some hints at security and confidentiality—a bizarre reason considering the security clearances held by contestability staff. It appears that some mandarins simply didn’t want it.

Despite the review to end all reviews deliberately ensuring Defence’s leadership had an appropriate in-house contestability function ‘to provide the best advice so that the outcome is delivered in the most cost-effective and efficient manner’, the Department chose not to use it.

And that’s the problem. When it’s the prerogative of mandarins to choose how and when they will seek contesting advice, they will avoid it if it could threaten their pet policies and programs, and if it is not possible to avoid it, they will simply ignore its analysis and advice. I would suggest that is how we ended up with a solution to Australia’s future frigate requirement that did not meet the Government’s fundamental requirements for a ship that was mature, in the water and in service, resulting in a grotesque program that might, after 20 years, deliver three warships at a cost of $27 billion.

Similarly, in all the autopsies of how AUKUS came about, I have not seen anything that suggests the Government or the mandarins sought contesting views. When the solution is already known, contestability is simply an annoying nuisance. Had contestability been performed, the bottlenecks in US submarine production would have been identified, along with the fact that they cannot be remediated before the US Navy has to hand over submarines to Australia under the so-called ‘optimal pathway’.

There are alternatives to leaving things up to the dubious discretion of the mandarins. For example, contestability is much more rigorous in the United States’ system, both internal and external. One key reason for that is because the US government has to publicly release much more information. Congress jealously guards its prerogatives, regardless of who is in the White House, so it passes legislation requiring the Administration to publish a vast range of information. Everybody examining the Australian Department of Defence’s programs knows that the best place to find accurate information on them is in US Defense Department disclosures, which provide much more granular detail than we ever get from our own officials.

Or consider the excellent work performed by US agencies whose existence and independence is guaranteed by legislation such as the Congressional Research Service, the Congressional Budget Office and the Government Accountability Office. Certainly we have the excellent work of the Australian National Audit Office, but it essentially backward looking, examining the wreckage of Defence’s acquisition train smashes. Compare that to the work of the American CBO, which examines the future achievability of government programs, for example in its analysis of the US Navy’s fantastical shipbuilding plans.

Here in contrast, parliament is controlled by the Government, which has no interest in disclosing more information—by definition anything useful released to the public is available to the Opposition, the last thing any government wants. So parliament doesn’t legislate for disclosure and transparency. In Australia, without clear legislated requirements, the Defence mandarins have incrementally reduced the amount of information they disclose, falling back on the usual justifications of security, commercial sensitivities, personal privacy and so on.

Tellingly, on virtually the same day the Government released the Varghese review, the Australian National Audit Office’s latest Major Projects Report was tabled in Parliament. It noted that on security grounds the Secretary of Defence had chosen not to disclose key information on 20 of the 22 major projects covered due ‘Australia’s strategic circumstances [having] markedly changed’.

This information, which had been provided in past editions of this reports, largely related to “initial operating capability” and “final operating capability dates” – critical information to assess the timeliness of equipment projects. Without this information, the ANAO was unable to assess the projects’ performance in key areas such as schedule. One of the most important pieces of external scrutiny has been cut off at the knees. In essence, the Australian public can no longer know whether the hundreds of billions of dollars of capability they are paying for is actually being delivered on a schedule that is relevant to our worsening strategic circumstances.

Without reliable information, contestability is hard to perform. That is the mandarins’ playbook: they tell ministers that external commentators don’t know what they are talking about, so their bleatings should best be ignored. We should simply trust the mandarins.

Of course, the mandarins do like some contesting views when governments pursue policies that the Public Service doesn’t like. I have no doubt many mandarins are aghast at the cost of AUKUS and the lack of proper policy development that generated the concept (for one, most of the mandarins were kept as much in the dark as the Labor Party was!). Some mandarins, at least those outside of the national security space, still harbour desires for Australia having a more autonomous foreign policy and a self-reliant defence force. So they don’t mind having some voices in the broader strategic policy community raising questions about AUKUS since it means they can remain silent and not put their million-dollar remuneration packages at risk or cut themselves off from the endless post-retirement gravy train of independent reviews and board memberships.

But there are strict limits to those voices. By definition, the mandarins are about ensuring discourse stays within the Overton window – a phrase coined by American analyst Joseph Overton defining the spectrum of ideas that politicians regard as acceptable to the general public. Keeping a steady hand on the tiller of delivering policies that fit the Overton window is how the mandarins justify their salaries. Proposing ideas outside the Overton window is by definition biting the hand that feeds the public service. This is of course results in perpetuating business as usual.

It’s the role of external thinkers, commentators, and folk that don’t succumb to the cozy consensus of Canberra’s senior bureaucrats to push the window wider. That of course means they have to advocate things that are by definition radical, unpopular, politically unthinkable, or worst of all, even new. This means that the motivations of the mandarins and of the ‘external’ national security community are fundamentally different. This is why Peter Varghese’s project of imposing SCNS coordination will not deliver true contestability of ideas and policies and will close the window even further on independent thinking.

It’s not surprising that the Varghese review’s recommendations are par for the course. Thay are aimed at ensuring we have the look of contestability for appearances sake, but not enough to make a difference. This is the outcome the review’s terms of reference were set up to deliver.

We’ve seen this many times before. Recently the Defence Strategic Review said that Defence’s innovation system is broken and needs to be fixed. But then ignoring every lesson and precedent that show fundamental defence innovation has to be conducted outside of the hierarchy and structures of defence organisations, the Government decides to keep the ‘new’ defence innovation system inside the Department of Defence. Bizarrely the Government and Defence boast that this means innovation will be fostered by being subject to the rigorous scrutiny and coordination of the Defence mandarins. The Varghese review is simply the same philosophy: more supervision by the mandarins will create the right kind of good ideas.

We should all be sceptical about the efficacy of a system of contestability that involves more oversight by the mandarins. If you don’t accept fancy concepts such as the Overton window, then simply look at the mandarins’ dismal track record of developing and implementing policies to provide an affordable housing market, achieve a timely transition to clean energy, ensure a supply of affordable gas, set manageable immigration intakes, build the infrastructure that meets that growing populations’ requirements, build an economy that can withstand downturns in Chinese demand for minerals, reestablish a viable domestic manufacturing industry, deliver defence capabilities on time and budget, and recruit sufficient numbers of people to fill the ADF—and that’s before we get to Robodebt. Let’s face it, the mandarins need all the help they can get in the ideas space.

At a time when we face not just growing strategic competition but also growing popular dissatisfaction with Canberra’s ‘business as usual’, the Varghese review offers a quaint form of nostalgia for an earlier age where the mandarins knew what was best for us, with the Overton window kept only slightly ajar. The best that can be said of Varghese’s work is that it is mostly harmless because it is largely irrelevant. But we are at a point in history where the window of new policy thinking needs to be thrown open. Quite frankly, the mandarins’ steering of the good ship Australia has failed us; they are the last people who should be defining what amounts to good and bad policy ideas.