^[1]

According to the newspapers, the government’s Competitive Evaluation Process (CEP) for the Collins replacement yielded bids of between $10 and $12 billion ^[2] for eight boats, with the price for 12 boats below $15 billion ^[3]. Delighted by the news, the South Australian government has renewed its call ^[4] for (the originally planned) 12 submarines to be built in South Australia. Who cares if we need them or not, they’re cheaper by the dozen.

I like a bargain as much as the next guy, but something doesn’t seem quite right. The Collins project spent $9.3 billion in today’s dollars to deliver six boats*. But the vessels had a problem or two, and extensive work had to be done to rectify what the SEA1000 website ^[5] calls ‘legacy defects’. That included $301 million to achieve an ‘interim minimum operating capability’, $361 million ‘to improve reliability, sustainability, safety and capability’ aboard the vessels, and $553 million to fit a combat system that actually worked. All up, the taxpayer was slugged $10.5 billion for six boats.

For the sake of comparison, let’s round the Collins price down to $10 billion—just in case some of the follow-on work included enhancements as well as remediation. Even then, $10–$12 billion for eight boats looks like a great deal compared with the $10 billion we paid for six. It means that the new boats will be cheaper than the old ones. And it’s not because we are seeking a less capable vessel; the Competitive Evaluation Process seeks a submarine with ‘range and endurance similar to the Collins’ and ‘sensor performance and stealth characteristics that are superior to the Collins’. Moreover, most observers anticipate that the new vessels will be larger than their predecessors. So why does it appear to be cheaper to build boats in the 2020s than it was in the 1990s?

Some of the apparent difference could come from the combat system being excluded from the future submarine cost estimate but included in the Collins cost. But we know from the Collins combat-system replacement program that an installed combat system costs only around $100 million per vessel, which is a small fraction of the overall picture.

Perhaps the foreign exchange rate has delivered savings on the import-dependent share of the project? Although figures of 70% local content are routinely bandied about, the import-dependent share will be much higher because many locally sourced components will depend on imported inputs and capital goods. A favourable shift in the exchange rate could, in theory at least, reduce the cost of the imported share. Alas no. The average AU$/US$ exchange rate ^[6] over the life of the Collins program (weighted by annual spending) was 73.6 cents—better than today’s 72 cents. As things stand, we’re actually more than 2% worse off.

The changing price of labour further deepens the puzzle. There’s no estimate available for the labour share of the Collins or future submarine projects, but it will be substantial—especially if the labour input to locally sourced components is included. For example, an indicative labour-share figure for a major warship is 50% according to Defence ^[7]. The ABS only began its present wage statistics series in late 1994 and the median spend on the Collins occurred in 1991-92, but extrapolating the average real annual growth (1.4%) in the reported data (1994 to 2015) yields a cumulative rise of just below 40% for the period 1992 to 2015. But the median spend on the new boats won’t occur until 2025 at the earliest, there’s another decade of wage growth to be added. Assuming (conservatively) that real wage growth moderates to 1% per annum, that still yields a 54% increase in unit labour costs between the Collins and future submarine projects.

In principle, it’s possible that the unit labour cost growth has been offset by productivity gains in manufacturing technology. Indeed, the automobile industry has managed to do so through mechanisation and supply chain refinement. But submarine yards don’t lend themselves to intensive mechanisation in the same way as automobiles; the production runs are too small to justify the investment. In any case, the government’s 2013 Future Submarine Industry Skills Plan envisaged a workforce of between 1,000 and 2,000 for the future submarine (excluding design, combat and platform system engineering staff), whereas the ASC workforce averaged only 821 between 1990 to 2000. So it appears that we’ll actually need additional rather than fewer people—so much for productivity gains.

Setting aside the details of labour demands and wages, the long-term historical trend in successive generations of military equipment has been for real unit costs to increase ^[8]. To the extent there are productivity gains, they are invisible in the data. Indicative rates of growth for advanced defence platforms typically run at 3-4% or more above inflation ^[9].

We are left with a paradox. Australian-built submarines are falling in price at the same time as they improve in quality and grow in size—despite a greater than 50% real growth in unit labour costs and countervailing historical trends in the cost of defence platforms. A cynic might observe that a non-binding beauty contest such as the CEP allows bidders to promise low prices and ratchet up the figure when the competition has left the field. Oops, I guess I’m a cynic.

*Unless otherwise stated, all historical figures have been converted to 2015-16 dollars using the CPI.