I was interested to see a new approach to submarine procurement pitched by Hugh White and Michael Keating ^[1] last week. They got quite ^[2] a bit ^[3] of ^[4] media ^[5] coverage ^[6], as tends to be the case for criticism of the future submarine project. (Though I’d be the last person to complain about that …)

The Insight Economics report ^[7] (PDF) they were fronting is well worth reading, even though I don’t agree with all of its conclusions. (Disclosure: I provided some comments on an earlier draft.) I don’t think it’s as easily dismissed as a ‘beat up’ as the defence minister’s response ^[8] suggests. The report is partly an economists’ view of the future submarine project, with technical considerations less prominent. As the authors put it:

If the issue is framed in terms of the cost-effectiveness of a very substantial public investment, rather than solely in the narrow framework of the technical capability of the various platforms under consideration, we believe that economic and public policy specialists can make a considerable contribution to defence procurement and complement the technical expertise deployed within Defence.

That approach is both a strength and a weakness. It’s a strength because it makes sense to treat defence like any other area of government spending and policymaking. A dollar spent on submarines can’t be spent on health or welfare. Experienced eyes looking at the problem from another angle can help explore the cost–benefit calculus. But it’s a weakness when an insufficient appreciation of technical matters clouds judgements of alternatives. I think the report has the right diagnosis of risk in the future submarine project, but it’s wrong in its prescription for a ‘Plan B’.

Let’s start with the diagnosis of risk, where I think Insight Economics is on the steadiest ground. The report classifies risk into four categories: strategic, economic, technical and industrial. In fact, they can’t really be considered in isolation. For example, the report says that ‘Australia is paying too much for twelve conventional submarines’. But ‘too much’ is a term that doesn’t have an absolute meaning. If the admittedly eye-watering $50 billion (or more) that the fleet will cost retires a commensurate or greater amount of strategic risk, then it’s a prudent investment—as a group of economists should appreciate.

I think the answer to that cost–benefit question comes out as a resounding ‘no’. Not because I think we don’t need submarines to face an uncertain maritime security problem—I think we do ^[9]. But the security problems we’re most worried about loom large now, and the government’s own experts think that they’ll come to a head between now and 2035—just about the time the first one or two submarines are entering service. It’s hard to argue with an assessment of unacceptable strategic risk in the submarine program. We could well end up paying ‘too much’ because our investment matures after the time that we really need the payoff.

As for technical risks, I think the report is also right to say that we’ve opted for a solution that pushes the boundaries of conventional submarine design. I was much happier when the ambition in the program seemed to have been wound back ^[10] to something achievable with less technological stretch. If they are realised, technical risks could further exacerbate the strategic risk through schedule slippages. And, as Augustine’s work tells us ^[11], extra percentage points of capability drive the cost risk upwards.

Finally, the industrial risk relates to the other three. Deciding to build the boats here drives up the cost risk because we’ll be starting at the top of the learning curve (let’s hope it’s a steep one ^[12]). And the more bespoke a project is, the greater the schedule risk ^[13], and hence the greater the strategic risk. So I agree that we’ve opted for a solution with higher than necessary risk. If it doesn’t work, we could be scrambling for an alternative.

The report’s preferred ‘Plan B’ is a rapid acquisition of ‘modified off-the-shelf’ submarines, which the authors argue would be a better use of resources than extending the life of type of the Collins. In that approach, the limited endurance of such boats (and of their crew) would be offset by operating from forward bases, with the support of a submarine tender vessel. It’s hard to see how that could work in a warlike setting. The whole point of wanting submarines in the waters of the 21st-century Asia–Pacific is that the surface is simply getting too dangerous—even for very well-armed warships ^[14]. A tender vessel represents a significant vulnerability and, even if it could be adequately protected, its presence is a sure-fire indication of submarine operations.

But let’s be generous and assume that we can operate a submarine tender close enough to a hostile power’s waters to get smaller submarines into patrol areas for useful periods. There’s still a problem of payload. As the operating environment becomes more difficult, and detection methods more sophisticated ^[15], submarines will have to be able to stand off at longer ranges and remotely deploy sensors and weapon systems. Unmanned vehicles of various types will most likely be the preferred way to produce an effect on an adversary, and being able to deploy them in large numbers will be important if a sustained surveillance and strike capability is required. Small submarines will thus be more limited in what they can do than larger ones that can act as ‘mother ships’. (In fairness, the planned Shortfin Barracuda seems to have many extra tons of payload available, so it scores well in that respect.)

For those reasons, the recommendation of off-the-shelf boats is the new report’s Achilles’ heel, and it’s not surprising that the government has jumped on that to dismiss it ^[16]. That’s a pity, because there’s otherwise a lot of sensible discussion in its pages. I’ll come back next time with my alternative Plan B, which builds on recent hard work on the Collins fleet. It’s not without problems too, but I think it does a much better job of matching risks and benefits.