Once again there’s a lively debate on the merits of having tanks (and armoured vehicles generally) in the inventory of the Australian Army. And once again both sides are simultaneously right and wrong. In fact, this debate is a good example of the lack of sophistication that’s all too common in the discussion of defence capability.

On one side, there are those who argue that tanks don’t add appreciable capability on the modern battlefield (at least those battlefields Australia should care about), that they are too heavy, that they are about to swarmed to death by cheap but lethal drones, or that long-range missiles are a more important addition to our land forces.

On the other side are those who argue that tanks are a must-have as an essential part of combined arms manoeuvre—usually adding that infantry and tanks work together to produce an effect that is greater than the sum of the parts. That camp points to Afghanistan as an example of a theatre in which the deployment of tanks could have helped protect Australian soldiers and made life more difficult for the insurgents being hunted.

The complication here is that all those things can be true at the same time, so it’s worth working through this example from first principles to see what the real questions should be.

The first principle is that adding more capability to the Australian Defence Force’s order of battle always adds more capability. That sounds like a tautology (because it is) but it’s still worth pointing out, because all too often the opponents of a weapon system acquisition simply write the capability boost off as inconsequential. But an army with armoured vehicles can do things that an army without them can’t, or at least not at the same level of risk. Armour is useful in circumstances in which armour is useful.

The second principle is that every capability has an opportunity cost. If the defence budget were infinite, we would simply give the ADF as much capability as we could acquire. But, in the real world, money spent on armour can’t be spent on submarines, and vice versa. Choices must be made. And it’s not just money; every new capability (which includes not just the hardware but also the people, training, doctrine development, support arrangements etc.) requires skilled and experienced people to implement—and those are a scarce resource as well.

The third principle is that defence capability should be acquired to retire risks to our national interests, and must do so commensurate with the cost of acquiring it. The ADF is not an end in itself. Rather, it is one of the tools the government can use to help shape the international environment, manage crises or—worst case—fight a hostile adversary.

A corollary to the principle is that no ADF capability is worth having for its own sake, or at any cost. Having a submarine capability, an air combat capability or a combined arms team is useful only as long as (a) its cost is less than the potential cost to the nation of adverse events it could prevent or ameliorate and (b) there isn’t a more effective use of the resources needed to acquire and sustain it.

The ‘capability calculus’ for any acquisition decision requires an estimate of cost-effectiveness to be consistent with the three principles above. That’s where things get difficult. A mathematician would estimate the expected costs of strategic risks by multiplying their probabilities by their likely impact if they are realised (the ‘expectation value’) and compare that to the cost of the capability acquired to address those risks. The trouble is that we don’t have a way to precisely estimate probabilities—history is remarkable contingent and thus provides little guide—and we don’t usually know the extent to which our capabilities will be successful in retiring the risks anyway.

In practice, to the extent to which they are done at all, cost–benefit estimates are based on gut feelings and military experience. Or, even worse, hands are thrown up in the air and the future declared unknowable—thus justifying any new capability on the grounds that it could be useful in some circumstances. That’s pretty much how we keep coming back to a ‘balanced force’, regardless of changes in technology or our strategic circumstances.

It would be a useful exercise to force anyone advocating for a pet project to state their estimates of probabilities and costs. While precision is not possible, it would at least force underlying assumptions and prejudices out into the open where they could be debated.

Applying all that to the tanks/armour debate, we come down to the following mathematician’s question: is the total cost of acquiring a modernised armoured capability less than (take a deep breath here) the sum of the products of the probabilities of scenarios in which armour would be useful, and the additional costs that Australia would incur in each scenario if we didn’t have armour available?

To make that more intelligible, imagine a situation in which we only have to worry about two possibilities: another Afghanistan-like mission or a Pacific war between major powers that Australia is drawn into. We would have to estimate a probability of each occurring, and then estimate how much extra cost we would incur from engaging in each without the proposed future armoured capability.

To decide for or against procuring armour, we have to calculate the expectation value of that future cost and compare it with the price tag for the proposed acquisitions. If you are doing anything other than that when arguing capability, you are either guessing or arguing on a basis other than cost-effectiveness. In practice, far too many decisions are based on preferences rather than probabilities and weighted judgements.

I suppose I have to give my estimates of value at this point. I’ll preface it by saying I wasn’t one of those criticising the initial acquisition of the Abrams tanks back in 2004. We got them at fire-sale prices, paying only for an upgrade before delivery, not for the tanks themselves—the sticker price was around $7 million per tank. At that price the boost to the army’s capability looked to be pretty good value, and the opportunity cost was low—maybe half a dozen fast jets.

Looking at the current proposal, I come to a different conclusion.

I think another Middle East deployment is more likely than a US–China war in the next 10 years—perhaps five times more likely. But the cost to Australia’s interests of losing in the Middle East is low. We’ve effectively lost a couple of times there in the past 20 years with only minor strategic impacts (which is not to downplay the loss of soldiers’ lives). A lost Pacific war could be existential, while another muddled outcome in the Middle East wouldn’t be, so it’s hard to assign a relative value.

But let’s say the costs are 50 times greater. Since the $2.9 billion price tag for the new-generation tanks would probably pay for 20 additional F-35s or additional force multipliers such as air-to-air refuellers, P-8s or Wedgetails, all of which would be much more efficacious than armour in a Pacific war, that’s where I’d be putting our money this time.