The recent defence strategic review ^[1] has much to recommend it. Even the publicly released unclassified version is the boldest statement of Australian strategic policy in many decades. But it is not without flaws. The chapter on deterrence (Chapter 4), for example, is either obscure or vague on a range of issues. Given that the entire chapter consists of 23 sentences, the level of obscurantism in relation to aggregate length is impressively high.

True, in the land of deterrence, a degree of ambiguity is, often, no bad thing. And the DSR’s predecessor, the 2020 defence strategic update ^[2], contained a single paragraph, of just three sentences, which implied a wholesale shift in Australian thinking on extended nuclear deterrence—one reason for reading this year’s review so closely on that point. In any event, deterrence is meant to be a key focal point of this review, so let’s try to untangle what we can.

Reading through paragraphs 4.1 to 4.13, the problems begin almost immediately. Paragraph 4.1 states that ‘deterrence is about compelling an actor to abandon or defer a planned strategy’. Frankly, it would have been better if the authors had not defined ‘deterrence’ as a subset of ‘compellence’ in the first four words of the chapter. The classic deterrence theorists tend to underline the differences between the two: deterrence is about stopping an intended action; compellence about forcing an intended action. Deterrence maintains the status quo; compellence changes the status quo.

In practice, of course, the world is seldom so neat. What Putin’s doing in Ukraine—attempting to force the country to abandon its independence—is compellence. He backstops conventional aggression with nuclear threats to keep NATO from interfering. On face value, those threats are intended to maintain the status quo in regard to the European nuclear balance, but are deployed to help undo the status quo in regard to an independent Ukraine.

It would be unreasonable to expect the review’s writing team to wrestle with the finer points of deterrence theory; that’s not its purpose. But the casual use of the compellence framework here suggests that Australia’s view of deterrence bears some similarities to Putin’s view—which is untrue.

Paragraph 4.3 says—accurately—that deterrence happens inside the head of the adversary’s decision-maker, so deterrent threats need to be credible. Moreover, says the review, we don’t know it’s working. Granted, deterrence works best on risk-averse decision-makers. It makes cautious leaders more cautious. But it seems to work well enough even in relation to more brazen leaders. The test of whether it works can’t be whether or not certain events happen in the real world—because we can’t attribute deterrence as the actual cause of a non-event. But we should look for evidence we can measure, in particular an adversary more cautious in its decision-making.

And that brings us to the nub of the problem: the pattern of thoughts straddling paragraphs 4.6, 4.7, 4.8 and 4.9. The first two assert a historical practice of deterrence by denial—courtesy of the defence of Australia doctrine—but accept that this model of deterrence was designed for less challenging times and is no longer suitable for today’s regional environment.

The plot thickens in 4.8: in future we will need to develop a capability to unilaterally deter any single country from attacking Australian forces or territory. The single-country test is new. And so is the deterrence requirement: we have to be able to deter unilaterally. That is, another country is thinking about attacking us and no one’s coming to our aid.

This paragraph is susceptible to a range of interpretations. Pessimists naturally presuppose that the single country we might have to deter unilaterally is China. If that were to be the case, Australia might well need an independent nuclear arsenal. But is China the prime candidate? It’s comparatively easy to imagine China and Australia coming to blows as part of a wider regional conflict. But, putting the matter brutally, Australia isn’t important enough for China to attack unilaterally. I read this paragraph more optimistically, as setting a fairly low threshold for what we might have to manage courtesy of unilateral deterrence.

That interpretation fits best with paragraph 4.9, which says the appropriate test for whether or not we can successfully deter any single country turns on whether or not we can prevail in the northern maritime approaches to Australia. That’s problematic since the review has already declared such an approach ineffective against higher-level threats.

The flow of thoughts across those paragraphs is complex and contradictory. One can hope that the classified version contained a better-reasoned argument on this point but, if so, that sense has been lost in the process of the thousand cuts. I think the best way to approach those paragraphs is to see them as an answer to the questions raised by the 2020 update. In short, how does Australia see the future of extended nuclear deterrence? And what did Australia mean when it asserted a claim to have more of the instruments of deterrence in its own hands?

On the first question, the DSR is not particularly helpful. Paragraph 4.10 states that our best protection against nuclear escalation is US extended nuclear deterrence ‘and new forms of arms control’. But the review is silent about the future of extended nuclear deterrence—indeed, over whether it has a future—and silent too about those arms control agreements. It appears to have no good understanding of the transformative effects of developments currently under way in the nuclear field—the rise of a tripolar nuclear world, the emergence of more risk-tolerant leaders within the traditionally responsible P5, and the spread of ICBM capabilities to a rogue state—upon both assurance and regulation.

There’s some danger 4.10 will be read as an endorsement of the Treaty on the Prohibition of Nuclear Weapons ^[3], or at least of a TPNW which has solved the problems of verification, enforcement and universality that currently make it unsuitable as a vehicle for nuclear disarmament.

Similarly, the single-country test may be an early marker of how much deterrent power the Australian government wishes to have in its hands, but it’s an imperfect indicator which needs further work.

The government confronts a difficult agenda on both deterrence and arms control—one made more challenging by AUKUS, which is sucking the oxygen out of other nuclear issues. Unhelpfully, government ministers have been drawn to a mantra of ‘nuclear propulsion good, nuclear weapons bad’, which has hindered rather than helped good thinking on extended deterrence. While US allies in northeast Asia—South Korea in particular—have energetically pursued greater nuclear engagement, Australia has been left back at the starting line.