^[1]I listened last week as my colleague Mark Thomson launched this year’s budget brief ^[2]. One of Mark’s points was that spending on any defence related proposal should be weighed against the gains in security that’ll accrue as a result. He worries that a surfeit of funds for defence might see proposals of ‘progressively diminishing merit’ brought forward. He’d also agree that it’s possible to underfund defence, in the sense of having unretired security risks that, if realised, would cost more than the steps required to mitigate them.

A good principle for any expenditure of any public funds is to spend whenever net benefits exceed the costs, but not otherwise. But I started wondering whether the application of cost-benefit analysis in a ‘pure’ dollar-based form to defence spending is (a) possible and (b) likely. I think the answer is ‘only sometimes’ on the first count and ‘probably not’ on the second.

Let’s start with ‘possible’. Sometimes a dollar cost is calculable, at least approximately. Examples might include the forced shutting down or destruction of Australian mines or resource extraction plants, or a disruption to shipping trade lines. The estimated cost of mitigating those risks through defence investment (again necessarily inexact) can then be balanced against the potential loss. Even the lives of Australians can be valued for this sort of calculation. That might sound distasteful, but it’s the basis of decisions about road engineering for safety upgrades, and for how courts set compensation.

Defence spending isn’t like buying an insurance policy ^[3], although that’s one of the benefits. In many cases, defence spending also provides an ability to shape the strategic environment, either through deterrence or military engagement, which has the net effect of reducing the likelihood of bad events occurring. It’s the sum total of benefits that have to be weighed against the costs.

And examples of strategic risks for which the costs or the benefits are readily quantifiable are relatively rare. Thucydides famously identified three major motivations ^[4] for entering a conflict: fear, honour and interest (or ‘treasure’). Where either gaining or protecting treasure is the aim, bean counters have some chance of being able to work out whether it’s worth doing. But for honour or fear, forget it. A couple of contemporary examples make the point. After 9/11, the US launched two major regional wars, motivated in no small way by fear of further attacks rather than a cold actuarial calculation. We can argue about the ratio of benefits to costs in Afghanistan, but I don’t think that anyone seriously argues that the Iraq War was worthwhile—which tells us that cost-benefit can be applied in a wider sense.

For an honour based example, look at the Senkaku/Diaoyu Islands dispute. Japan and China are locked in an intense battle of wills over some entirely unremarkable rocks. Even on the most favourable resources assessments, the potential dollar value is completely overmatched by the costs of a conflict between the two claimants. In this case honour is the balancing factor to financial cost. (See also ‘Falklands War’.)

Even when treasure is the major motivator and likelihoods can be estimated with reasonable accuracy, there’s still no guarantee that sound decisions will be made. Most people aren’t natural actuaries, and make poor choices when the stakes are high and emotions are engaged. As an illustration, let’s turn to that epitome of educational television, Deal or No Deal ^[5]. Watching it makes me despair; at every step there’s enough information available for contestants to pursue an optimal strategy. And at almost every step they don’t (and when they do, it’s usually by accident). It drives me nuts.

Let me explain how to play Deal or No Deal optimally. (You can read the rules and play an online version here ^[6].) Whenever a ‘bank offer’ is made, add up the remaining sums available and divide by the number of suitcases left. If the number is bigger than the bank offer, play on. If it’s smaller, take the offer and leave. That’s it. As with many games shows, a mathematician could take a ruthlessly logical approach and optimise their outcome but there’d be no drama, no excitement and no tears. That’s why mathematicians don’t often appear on games shows. (There are probably other reasons too, but let’s not go there.)  Mathematicians also aren’t generally invited to meetings to discuss national security.

Finally, there are political, commercial and professional interests at play in strategic decision making. I’ve written before about the professional motivations of the Services ^[7], and the potentially distorting effects of commercial interests are obvious. But politically, governments must be seen to be capable of protecting the nation and its people, not just physically but also in the sense of national honour (Thucydides again). For the starkest example of how that translates into strategy on a grand scale, consider French General Ferdinand Foch ^[8]’s response to a British query regarding the smallest useful force that could be provided to support France in the event of a war with Germany. His answer ^[9] was: ‘one single private soldier and we would take good care that he was killed’.

Andrew Davies is senior analyst for defence capability and director of research at ASPI. He doesn’t buy lottery tickets.