Articles by " Andrew Davies"

Edward Snowden, the media and the Pulitzer

The Washington Post

The decision last week to award a Pulitzer Prize to the Guardian and Washington Post newspapers for their coverage of classified material leaked by Edward Snowden has refocused attention on the pros and cons of both Snowden’s and the newspapers’ actions.

Some have praised the decision and have hailed the newspapers for being both ‘judicious and brave’ in their handling of the material. Others, including one of my ASPI colleagues, see little value in awarding the prize for what amounts to an unauthorised release of state secrets.

In truth, there’s merit in both positions. Unlike Wikileaks before it, which largely released material that was embarrassing to governments and militaries but has been of little lasting security harm, the Snowden case involves extremely sensitive material that has the potential to cause deep and lasting harm to the ability of America’s intelligence agencies—and, because of the five eyes relationship, Australia’s—to perform their roles. Making public some of the access points for interception of material, and the technological tricks required to exploit it, will play to the advantage of those trying to keep their communications out of the hands of American and allied agencies. And the nature of the intelligence business is that it’ll be difficult to know what’s been lost—it’s hard to quantify intercepts that don’t happen.

In some cases that won’t be such a bad thing. I’ve argued before that intelligence collection against European allies didn’t pass the cost–benefit test that should’ve been applied. Losing those information channels probably won’t do the West’s security interests any harm. But taking a black and white view and painting American intelligence as the bad guys in all this makes little sense. In some cases, such as Snowden’s release of information about methods used to intercept al-Qaeda communications in their Mosul network, that clearly isn’t the case. Some of those aided by Snowden are clearly in the ‘black hat’ camp by any reasonable measure, and both terrorist groups and authoritarian states will benefit from knowing the ‘tricks of the trade’ used by the US and its allies.

The fact that Snowden himself is now ensconced in Russia and is offering himself up as an asker of Dorothy Dixers to Mr Putin suggests that he’s somehow managed some prodigious mental gymnastics of self-justification. That said, some good will also come of Snowden’s actions. Some of the NSA’s activities that were disclosed weren’t consistent with stated practice. The subsequent investigations by the press and civil liberty groups have uncovered material that makes the initial release reasonably fall into the whistle-blowing category. It seems clear that oversight wasn’t working as designed, and that the NSA wasn’t even working especially cooperatively within that system. The New York Times reported on a ruling from the Foreign Intelligence Surveillance Court—a body set up to oversee the NSA’s collection activities on US soil. It was sharply critical of both the NSA’s ‘repeatedly inaccurate statements’ to the court and of some of its collection and analysis activities, which it judged to be unconstitutional and misleadingly reported to oversight bodies.

It’s hard to argue that it’s not in the public interest to have an open discussion of the matters raised in the NY Times article. That’s especially the case when the efficacy of the oversight mechanism itself is called into question by the judge charged with administering it:

“Contrary to the government’s repeated assurances, N.S.A. had been routinely running queries of the metadata using querying terms that did not meet the standard for querying,” Judge Bates recounted. He cited a 2009 ruling that concluded that the requirement had been “so frequently and systematically violated that it can fairly be said that this critical element of the overall … regime has never functioned effectively.”

Similarly, it’s a principle of good government that the expenditure of taxpayer’s funds should be done as transparently as possible. The post-9/11 ‘black hole’ that was the US intelligence budget violated that principle for no good purpose. It’s hard to see how the Washington Post imperilled anyone’s security by a selective release of the FY2013 budget papers. For example, knowing that 33% of the budget was going to countering violent extremism while only 8% was being spent on enhancing cybersecurity can only help inform a debate on the relative priority of those issues.

When assessing Snowden’s behaviour, the good outcomes (which are relatively easy to identify) have to be weighed against the bad (which mightn’t be so obvious) and I think he’s done more harm than good. But I can’t be similarly critical of the press. A robust liberal democracy depends on both an informed population and a fractious and difficult press that’s prepared to follow the dictum that ‘news is something that someone doesn’t want printed ‘. That’s certainly been the case here.

Andrew Davies is senior analyst for defence capability and director of research at ASPI. Image courtesy of Flickr user Max Borge.

Submarines by the dozen?


The press has made much of a perceived backing down from a plan to build 12 submarines. We say ‘perceived’ because no-one has actually said that. But it’s true there’s been some very careful language choices around submarine numbers, including by the Defence Minister, here in his speech:

… my primary focus is not on numbers but on the capability and availability of boats required to meet the tasks set by government.

And here in a doorstop:

To [put] a number on submarines is a distraction. What we want is a long term capability that can be sustained as an enterprise, as an asset that can go long into the future building submarines.

Read more

It’s true that the focus ought to be on the delivery of defence capability that’s well-matched to strategy and to the budget. But talk at the conference, both from government and from the bureaucracy, has been about the need for an enduring industrial capability for submarines. While no-one has said so in as many words, an enduring submarine design and build capability all but mandates moving to a continuous build program—the case for which was laid out in detail in DMO’s Future Submarine Industry Skills Plan last year. Another criterion is avoiding any capability gap that might otherwise occur at the end of the Collins-class lifetime.

We think meeting both of those criteria is only really feasible if the fleet constitutes around 12 boats. We’ll explain why below, but first observe that France has 10 boats (and exports others), and still has management challenges in keeping its industrial capability intact. The UK’s fleet of 11 submarines (and no exports) has barely provided enough continuous work. By general consensus, Japan sustains its industrial submarine capacity pretty well through a rolling production model, but it has 16 in service and is expanding to over 20. Maybe we could come up with a model that works with fewer than 12 boats, but clearly we’d have our work cut out.

Other numbers here aren’t especially promising either. As we pointed out in our 2012 Mind the gap paper (PDF), a Collins life extension will take them out to 2030 (and beyond that for a few boats). By then they’ll be around 30 years old—not unusual for naval platforms.

The recent success in improving Collins sustainability has seen a move to a ’10 years on, 2 years in maintenance’ operating cycle (it was previously 8 + 2), allowing more efficient use of those expensive assets. Future boats could thus serve for 22 years as a minimum, and 34 if they do three cycles as the Collins class will come close to.

So if we had 12 submarines and kept them for the minimum 22 years, we’d need a new one every couple of years. If the number fell below 12, we’d have to slow down further, raising the question of what constitutes an efficient use of the investment required to sustain shipyard and design capacity. Such a slow production rate wouldn’t replace the six Collins boats in the right timeframe; we’d have to produce a batch of four to six fairly quickly and then slow down—but then it’d be hard to avoid having 10–12 boats at some stage.

Of course, 22 years is a remarkably short life-of-type for a submarine. So the question is whether the benefits of an ‘enduring capability’ justify the additional cost of replacing vessels more frequently?

A quick estimate isn’t encouraging; moving from a 34 to 22-year lifespan increases the capital cost of maintaining the fleet by more than 50%. Even with potentially higher productivity and potential savings from avoiding mid-life upgrades, it’s likely there’d be a substantial cost premium. Then there’d be the added costs of maintaining administrative and managerial overheads continuously, within both industry and Defence.

A continuous build program of ships and submarines would also lock the government into maintaining the size of the submarine and maybe surface fleets. Navy might see that as an added benefit; no need to make the case for the next generation at replacement time. But from a broader defence perspective it would fix the minimum size of a large and expensive part of the force structure. And from a public policy perspective it would lock in a substantial chunk of what was previously discretionary spending.

Call it what you want—an enduring capability or a continuous build program—it means that we’d be creating either a private or publicly-owned monopoly submarine production entity. As the bad old days of government-owned shipyards demonstrated, ensuring productivity from a monopoly supplier is a far from easy task.

Finally, it’s worth noting that the main rationale for an enduring capability is a desire to meet some ambitious and uniquely Australian requirements—the prime source of cost and schedule overruns in other defence equipment over the years.

Andrew Davies is senior analyst for defence capability and director of research at ASPI and Mark Thomson is senior analyst for defence economics at ASPI. Image courtesy of Flickr user sir chalky.

Reporting some good news

Chief of Navy complained at the conference today the press loves to write about ‘dud subs’ but isn’t as keen on good news. And there’s some good news to report today, in the form of the latest instalment of the Coles review into the availability of the Collins class boats.

This is the latest in a series of reports, which have been tracking the progress of work to turn what was a moribund fleet into the military capability it was originally designed to be. We’ve been tracking progress here on The Strategist as well, from the dark days of poor management and outcomes through the noticeably better but still qualified performance noted in the previous report.

And the dark days were dark indeed—the new report reveals that Australia’s submarine capability basically collapsed in the second half of 2009. There were no days on which three boats were even in principle available for operations, and two boats were available less than 10% of the time. That compares to Navy’s targets of:

two deployable submarines consistently available, with four submarines available to the Fleet Commander and of these four, three submarines consistently available for tasking with one in shorter term maintenance and two submarines in long term maintenance and upgrade.

Read more

But that was then and this is now. The figure below shows the fleet availability through to the end of last year, compared to the international benchmark. (The absolute figure isn’t given, but our previous estimates were about 1,200 annual days from the fleet of six boats.) Clearly things are on the mend. Other stats in the report show that maintenance times are coming down, and defects are less frequent than before. We’re not up to the benchmark yet, but are well above the lowest point, and even ahead of where we hoped to be.

Source: Collins Class Submarine Sustainment Study, Department of Defence, p. II

Source: Collins Class Submarine Sustainment Study, Department of Defence, p. II

But, just as before, this good news comes with some qualifications. Some obsolescence issues are still to be addressed, and the Collins life extension program that we’re now committed to will need to address those. And there’s a message in here for government as it contemplates DMO’s workforce. A shortage of qualified engineers in that organisation is identified as a potential threat to further progress. Coles observes that DMO doesn’t have the flexibility to manage its workforce appropriately, which suggests that some targeted reforms, as well as apparently imminent cuts, are needed in the acquisition organisation.

Andrew Davies is senior analyst for defence capability and director of research at ASPI.

Submarines: does Moore mean less?

Back when I were a lad, some 40 years ago, my family was just thinking about swapping our trusty old black and white TV for the technological marvel that was colour. The internet was yet to escape the US DoD’s clutches—not that anybody had a computer at home in any case—and the first live cricket telecasts from England were a big deal. Back then a high-tech car had seatbelts, not computers.

The reason I’m getting all nostalgic is because I’ve been putting together my thoughts for ASPI’s Submarine Choice conference. I was pondering the fact that, if past platform lifetimes are any guide, the future submarine will still be with us 40 years from now.

I’ve written before about the exponential growth trend in computer power known as Moore’s Law. The past 40 years has seen an increase in computing power of over a million times. And, unlike the cost of military hardware, costs have trended steeply downwards as performance has soared. The next 40 years will likely see the same increase in our ability to capture, process and move information around. Read more

If you say that fast, it seems comprehensible enough, but what we’re talking about isn’t the same increase as we’ve just experienced, it’s a million times that again. Computing power in 2054 will a trillion times what was available in 1974. A 4G smart phone with its data capture, storage, processing and access capabilities will be as quaint an antiquity in 2054 as it was unimaginable in 1974.

It’s important to try to understand what that means for a $40 billion project. We like submarines because they have a formidable power projection capability, and can venture into even highly contested spaces, where more visible platforms can’t. If that was significantly compromised, then we’d be either looking elsewhere with our investment dollars, or at the very least rethinking how we’d design and employ submarines.

Although the past few decades have seen submarines become more effective rather than less, I think there’s a good chance that they’ll lose that stealthy edge in the future. They’ve had it so good up to now due to advances in radiated noise management, coatings and hull designs to make detection by sonar more difficult. They’ve got ahead of the detector capabilities.

But ultimately there are physical limits to what can be achieved and we’re getting into diminishing returns. Further improvements are possible, but they’ll be smaller than past ones, and likely cost more. And on the detection side, Moore’s Law is going to bring hugely increased processing power so that sorting even weak signals from noise will become faster and more reliable and hiding that much harder.

Recent submarine classes exhibit relatively modest improvements in noise management compared to previous generations. (Source: US Office for Naval Intelligence data, via Wikimedia Commons)

Recent submarine classes exhibit relatively modest improvements in noise management compared to previous generations. (Source: US Office for Naval Intelligence data, via Wikimedia Commons)

The other relevant trend is the development of unmanned platforms: small and relatively uncomplicated drones could be used to collect information and forward it to a powerful central processing hub. If the cost can be kept down, there could be hundreds or thousands of airborne, surface or subsurface drone-based sensor systems deployed in the choke points and littorals where diesel electric submarines are most effective. In my conference paper, I sketch some ways that might work.

That combination is likely to make sneaking large platforms into contested spaces prohibitively difficult. And it complicates life for conventional submarines even more than for nuclear boats because the littorals will be more dangerous than blue water.

I’m not the only one making those observations. USN chief ADM Greenert observed:

The rapid expansion of computing power… ushers in new sensors and methods that will make stealth and its advantages increasingly difficult to maintain above and below the water.

But technological advances tend to cut both ways, and the battle is often to the side with the right combination of technical capability and imagination. Greenert went on to say:

US forces can take advantage of those developments by employing long-range sensor, weapon, and unmanned-vehicle payloads instead of using only stealth platforms and shorter-range systems to reach targets.

Submarines will have to stand off from high-stake situations and exert their influence from a distance by deploying their own long-range remote or autonomous sensors and weapon systems. Rosie Turner’s piece earlier today suggests some future evolutions in unmanned underwater systems.

Those criteria pretty much rule out an ‘evolved Collins’, which probably can’t stretch to those requirements. That leaves us with two broad options:

  1. Go all out with the design of a large, fast, long-range boat that can operate at the highest level in a much more challenging future, or
  2. Temper our ambitions and settle for a fleet that will deliver value for money capability in less than the most challenging situations.

In other words, we have a really big decision to make right up front, and the stakes are pretty high. It’s going to be an interesting ride.

Andrew Davies is senior analyst for defence capability and director of research at ASPI. The full text of the conference talk is available here.

Changing of the guard

Changing of the guard

After almost two years as Executive Editor of The Strategist, it’s time for me to hand the reins over to my colleague Rod Lyon. With the team of Natalie Sambhi and Kristy Bryden in place to support Rod, I’m confident that this blog will be in good hands.

I’ll be putting my efforts towards my military capability studies and towards my contributions to the defence white paper ministerial advisory panel. And that’s the main reason for a change.

The white paper is going to be one of 2014′s major policy developments in areas of ASPI’s remit and it’s important that The Strategist continues to be a place where views on the content of the white paper can be aired and debated in as open a fashion as possible. If the experience of previous white papers is anything to go by, we’re talking about policies that will still be having a financial and capability impact decades from now. It’s important to get that right.

And I’m as confident in the ability of our readers to argue the merits of various approaches as I am in the editorial team to select and present them. I’m sure there are fun times ahead.

I’ll still be writing for The Strategist (sorry to disappoint), but on topics other than the white paper. Finally, thanks to the readers and contributors who’ve taken the time to share their views with me during my tenure. Your efforts were always valued.

Andrew Davies is now “only” senior analyst for defence capability and director of research at ASPI. Image courtesy of Flickr user jessicamulley.

What do the AWD problems tell us about the future submarine?

I recently took a look at the ANAO’s audit of the Air Warfare Destroyer (AWD) program. The report is a solid read at 320 pages, but should be required reading for anyone making decisions about the future submarine. And there should be a test afterwards.

The first, and most important lesson is the need for design stability—and for the design to be in a form readily digestible by those who need to turn it into hardware. A great many of the problems experienced in the AWD build have been due to either changes in design details, or the difficulty of translating the design drawings into concrete production activities.

The ANAO’s figure 5.7 (reproduced below) shows the number of entries in the project’s ‘problems and issue reports’ database by category. Nearly half the records related to design issues. The fault for those problems shouldn’t be laid solely at the feet of Navantia as the design house, although there have been attempts to do so. But it’s worth understanding what went wrong.

Figure 5.7: Categories of problem and issue reports, July 2009 - April 2013 Read more

Part of the problem is that this is the first time that Navantia has exported a design. They’re used to preparing drawings for its own workforce in its shipyards at Cádiz—a workforce familiar with both the design philosophy and the technical drawing equivalent of ‘shorthand’ employed and, more importantly, the tacit knowledge of ship production assumed by the designers. An inexperienced Australian workforce struggled to translate them into executable work. This problem reflects the lack of a shared understanding of design and build approaches—an issue that should’ve been identified during the long and costly analysis that lead to the selection of the Navantia design (of which the shipbuilder ASC was a party).

Exacerbating this problem, the Australian workforce lacked critical skills at the production supervision level. In the case of the well-publicised problems with the poor quality of the blocks built in Melbourne, Defence’s 2010 advice to the Minister is pretty clear:

… the poor build quality was largely the result of BAE Systems not having sufficient experienced production supervisors—workshop engineers and foremen—despite being one of Australia’s most experienced shipbuilding organisations.

BAE were having none of it, there were again attempts to lay the blame at Navantia’s feet, and we were treated to the unedifying spectacle of ASC and the DMO arguing the toss with BAE in the press. The ANAO makes it clear that there’s plenty of blame to go around.

Now all this might be OK if it was behind us, but it looks far too much like a taste of things to come. The future submarine project will face all of these problems and more. In the case of the AWD, we started a build with a first-time shipbuilder and an inexperienced workforce. Seven years on, productivity remains well behind planned levels and even further behind industry standards for an experienced yard—the result of a combination of overoptimistic planning and ineffective production management. That’s despite the benefit of starting with a design that was pretty well understood, with a similar vessel already in production in Spain.

The future submarine won’t be like that. If it’s to come anywhere near the articulated requirements, it won’t have much in common with boats elsewhere. If we stick with the expected European conventional submarine technology and an American combat system and weapons, we’ll have to import design elements from more than one supplier. In short, we’ll be starting a harder journey from no better a starting point than the AWD.

That’s not an insurmountable problem. Given time and resources, we could pick our way through the challenges, hire the experienced people needed to manage the project and do the due diligence required to stabilise requirements and, subsequently, the design. It’s having the time that’s likely to be the issue; much of the discussion about Australian naval shipbuilding has focussed on the so-called ‘valley of death’ facing the blue collar workforce as AWD work winds down. And that is a problem—workforce productivity takes time to build up to best practice standards. But we need to be careful not to embrace costly ‘make work’ projects with cost premiums that exceed the savings which continuity of work might deliver via higher productivity.

I’d argue that much more important at this stage is managing and harnessing the skills required at the white collar end of the process—the engineers, designers and production managers who have to sign off that the chosen submarine design is fit for purpose and ready for production. I fear that we’re already at the wrong end of the valley for many of those skills as the design work for the AWD recedes into the past. The ANAO estimates that the overall productivity in the AWD build is such that it’s costing $1.60 for every planned $1 worth of work. Their report suggests that a lot of that is to do with design and production issues, not just fabrication.

My high school metalwork teacher used to exhort us to work with ‘less haste and more speed’, which I think is exactly what’s needed for the future submarine. Here’s my dot point list of things we need to do:

  • ensure that senior engineering and design personnel from those firms chosen to provide major systems are embedded in the project team, not on the other side of the world
  • accept that we’ll need to import production techniques as well—which means that shipyard foremen and production supervisors should be imported with the design
  • ensure that we have a stable and suitably documented design, and resist the urge to get to the metal fabrication stage as soon as possible as a means of providing shipyard jobs.

Andrew Davies is senior analyst for defence capability at ASPI and executive editor of The Strategist.

Decision time for Australia’s F-35 plans

Lockheed Martin's F-35 assembly line at Fort WorthThe government looks set to spend somewhere between $8 and $10 billion on F-35 Joint Strike Fighters, which would then constitute the bulk of the RAAF’s air-combat capability for decades to come. The aircraft has had a difficult upbringing, but there are a few myths about it that we think are worth dispelling.

  1.       Prices are still climbing.

Actually, no. The F-35 is never going to be the bargain price fifth-gen strike fighter beloved of glossy brochures in the early 2000s, but the price is now trending in the right direction. The last annual production batches have come in at lower prices than the US budget estimates. Of course there’s always potential for a cost-escalation while risks remain in the development program but, as shown in Friday’s post, that’d be a departure from the recent trend. Estimates of support costs are also coming down. Read more

2.      The F-35′s development continues to be a problem.

Well, this is partly true. There are certainly challenges remaining—software development and integration, and fixes for structural cracks in the US Marine Corp’s F-35B (which won’t directly affect Australia’s A-variant purchase), for example. But the overhaul of the program which began in 2010 produced a timetable that’s mostly holding—certainly far better than previous performance.

3.      Professionals in the military are still concerned about the performance of the aircraft.

In talks with the USAF and RAAF, the noises we get about the performance of the F-35 are overwhelmingly positive. In the words of General Mark Welsh, Chief of Staff, US Air Force:

When a fifth generation fighter meets a fourth generation fighter—[the latter] dies. We can’t just dress up a fourth generation fighter as a fifth generation fighter; we need to get away from that conversation.

If anything, the problem is that enthusiasm from the services that will employ the F-35 is so strong that it’s difficult for them to hear the case for scenario-based planning which explores less capable acquisition options.

The picture we get is that after a shaky start and more than a little denial, the JSF program’s now starting to behave as promised. That said, the history isn’t great. Prices have risen sharply from the 2002 quote of US$40 million per aircraft flyaway (around US$55 million in 2019 dollars). That compares to the latest USAF budget figure of US$97 million, an increase of 76%. As we noted above, the actual price is likely to be a bit less than the budget figure and US$90 million is a reasonable estimate. And there are other costs; because the F-35 schedule has slipped, old aircraft need to be kept going until they can be replaced. But even so, the revised price is still within the AIR6000 budget, which seems to have contained a generous—but subsequently necessary—allowance for cost increases.

Similarly, schedules have blown out. But now it seems likely that the US Marines will be able to achieve Initial Operating Capability (IOC) with their F-35B short take-off vertical landing ‘jump jet’ variant sometime around the end of 2015. The F-35A conventional take-off and landing variant, which Australia is buying, should enter service with the US Air Force in the second half of 2016.

That gives Australia some margin of comfort for the RAAF’s own planned IOC of 2020, which would allow the phased retirement of the vintage Hornet fleet as planned. In fact, the first two Australian aircraft will be rolling off the production line in the States in the coming months, from where they’ll join the training fleet in the first instance.

There aren’t many credible alternatives. The only one would be for the government to decline an F-35 purchase, and go with a further purchase of Super Hornets, giving the RAAF a single-type fleet of strike fighters (augmented by Growler electronic warfare aircraft). That would provide some savings, at the cost of a lower overall capability than a mixed fleet including both F-35s and Super Hornets could offer. But new Super Hornets would have to be ordered now, as the production line is winding down, and spending an extra many billions of dollars out of the forward estimates period isn’t going to happen.

And there’d be other costs to backing-away from the JSF, not least the ripples it would cause in Washington—there’s alliance capital tied up in the F-35. On balance, and while past decisions might’ve been handled differently, we think the F-35′s now on a stable enough development footing that a decision to purchase the F-35 is both responsible and fits government priorities.

Our new report, out today, has more detail.

Disclosure: Lockheed Martin, the prime contractor on the F-35 program, and Boeing, manufacturer of the Super Hornet, are corporate sponsors of ASPI.

Andrew Davies is senior analyst for defence capability at ASPI and executive editor of The Strategist. Harry White is an analyst at ASPI. Image courtesy of US DoD Inspector General.

Graph of the week: F-35 update

The National Security Committee of cabinet will soon consider a submission from Defence regarding the proposed approval of a buy of (probably) 58 F-35 Joint Strike Fighters in addition to the 14 already approved.

We’ll be publishing a discussion paper on Monday that looks at the pros and cons of that proposition. And it won’t be a surprise that one of the major issues will be the state of the F-35 development and production program. The F-35 and the process of developing it have been subject to some trenchant criticism over the years—and with good reason. The performance of the prime contractor has left much to be desired at times, and the apparently hands-off management practiced by the Pentagon in the early years led to a series of cost increases and schedule slippages.

Some of those trends weren’t surprising, as early estimates of costs were well below the historical trend for combat aircraft. (The most recent ones certainly aren’t.) And the joint development of three very different variants—one for conventional takeoff and landing, one designed for carrier operations and a ‘jump jet’ for the US Marines—meant that initial schedules were seriously optimistic. Read more

But some of the underperformance could squarely be sheeted home to poor program management. Combined with a tendency of those involved in program management and execution to downplay all of the increasingly obvious problems, a serious credibility issue developed. Some of us got to the point where a comment from an F-35 spokesman that it was a nice day would require confirmation from the US Government Accountability Office.

That’s all part of what the now head of program Lt General Chris Bogdan  described here in Australia last week as the ‘tragic history’ of the F-35 program. But he also went on to describe a much tighter management approach today that has seen much improved performance against the program’s 2011 revised milestones. Let’s see what this year’s Pentagon budget papers—released last week—have to say about that.

Last year I presented some graphs of the history of US Air Force F-35 budget plans going back to 2006. They showed a pretty clear picture of the history as described above, but also hinted at a more stable present. This year’s figures reinforce that view.

The first chart (click to enlarge) shows the budget cost planning figures for each of the budgets since the 2008 financial year request (in March 2007). The three phases—I’ve labelled them ‘early optimism’, ‘reality intervenes’ and ‘recent stability’—are clearly visible. For aircraft purchased later this decade, as Australia’s planning to do, there’s been little change in the last three budget cycles. (All prices in 2014 US$ millions.)

F35 Graph 1

Source: USAF budget papers ‘aircraft procurement volume 1’

The second chart shows the planned cumulative production numbers of the Air Force variant. The same three phases are obvious.

F35 Graph 2

The same conclusion as last year applies—where the F-35 is concerned, the US Air Force is determined to put its money where its mouth is these days. Given the difficult budget environment in Washington, that’s no small vote of confidence.

Of course, there are factors other than program stability to be considered when deciding whether to spend billions of taxpayer’s dollars. We’ll come back to those next week.

Disclosure: Lockheed Martin, the prime contractor on the F-35 program, is a corporate sponsor of ASPI.

Andrew Davies is senior analyst for defence capability at ASPI and executive editor of The Strategist.

The AWDs and the auditors—round two

A computer generated image of the Air Warfare Destroyer (AWD). The project will deliver three world-class ships and their support systems to the Royal Australian Navy ( RAN).

When the Australian National Audit Office’s audit report on the Air Warfare Destroyer program was released last week, I was told by a veteran journo that it was a bit of a disappointment to his colleagues. They’d apparently been hoping for a story about a Defence managed fiasco that’d make for screaming headlines along the lines of ‘billions wasted says government watchdog’.

That’s an unfortunate reflection on the state of news reporting these days, because the ANAO report is a very important one, and it has implications far beyond the AWD program. It fills out much of the detail hinted at last year in ’round one’. And I’d have thought there are enough titbits in there to make for at least a moderately unhappy story if that’s what required to sell papers (or, increasingly, to draw clicks). But it’s also to my eye the auditor’s best Defence project effort to date, so let me at least put my vote of thanks on the record for an important piece of work.

This report is especially significant because there’s bipartisan support in Australia for continuing naval shipbuilding in country. So despite the shortcomings in the AWD build, more such projects are likely to follow—with costs to taxpayers running into tens of billions of dollars. It’s pretty important to understand what’s happened in this case. Read more

I’d recommend reading at least the summary of the report, but I’ll provide some comments on what I think are the most salient points. Let’s start with the headline issue of cost overruns: the alliance building the vessels is reporting an estimated final cost of some $302 million, or 6.8% above the target cost estimate. But given that the overrun in the 2012–13 financial year was $106.4 million as a result of increased labour, materials and subcontract costs, largely for the relatively straightforward industrial work of module construction, and that the difficult system integration parts are still to come, it could easily get worse than that. The ANAO says as much: ‘the current estimated cost [overrun] of $302 million … should be treated with caution; the cost increase is likely to be significantly greater. … the program is approaching the complex stage of systems integration when, historically, cost and schedule risks tend to rise’.

So by the time the program’s finished, we’re going to end up paying an even larger than expected premium for local construction of these vessels. The report tells us that the 2007 Treasury estimate of the premium was ‘around $1 billion’. My estimate was $1.5 billion (PDF), which probably won’t be far off the mark. We also learn that the ‘effective rate of assistance’—a measure of economic distortion due to the program—was estimated at 30%, a figure well above the level for other subsidised industries.

Those figures are sobering. Naval shipbuilding is an enterprise that’s consuming taxpayer’s resources that are therefore not available for other government programs. And there’s no shortage of lobby groups making the case for expanding it, arguing that it’s the stop-start nature of the business that has caused the problems seen here. Indeed, the ANAO includes some comments along those lines from two of the shipbuilders (whose performance fell short of what was expected when the project was approved):

… companies that have participated in the AWD Program, our company included, have faced a significant challenge from the need to re-establish capability, capacity and experience after the gap in naval shipbuilding that preceded the start of AWD construction.


… the costs of ramping back up to a competitive shipyard to maintain the indigenous shipbuilding capability have not been fully appreciated in terms of the magnitude of the investment required in facilities, recruitment, training and retention of the workforce to reach competitive productivity; once established, the shipbuilding capability will once again dilute and disappear if not utilised in an ongoing shipbuilding program out across the Defence portfolio.

As I said above, there’s bipartisan support to respond positively to these entreaties. So the policy question now becomes ‘how best to ensure that the local naval shipbuilding industry performs as well as it can’? The first thing to do is to try to work out how much of the underperformance of the AWD program is due to genuinely unavoidable learning issues in an industry that hadn’t built a warship in well over a decade, and how much is due to factors more amenable to best practice management.

There’s plenty of evidence in this report of start-up and inexperience woes, which seem to have taken pretty much everyone by surprise, despite $262 million being spent on due diligence activities before contract signature. And some of the problems should’ve been predictable. Looking over the report, devotees of Augustine’s and Gumley’s Laws will recognise some old friends. In paragraph 26, for example, we find:

… additional modifications primarily to accommodate the Australianised Combat System, which comprises an upgraded Aegis Weapon System and additional Australian elements to meet specific capability requirements. The selection of largely existing platform and combat system designs formed the basis of the DMO’s technical risk reduction strategy for the Program, and was considered at the time to provide a high level of comfort.

The ‘specific capability requirements’ might well be Augustine’s final 10% of capability, which we know leads to 1/3 of the cost and 2/3 of the risk. And Gumley’s laws remind us that combining off-the-shelf systems isn’t off-the-shelf. (MOTS + MOTS ≠ MOTS)

But that was then and this is now. And those errors, while regrettable, are now behind us. Hopefully we’ll learn from the experience and not repeat the beginner’s mistakes that have plagued this program. Next week I’ll come back to the report and tease out what it might teach us for future naval shipbuilding programs. There are some powerful lessons here—especially for the future submarine.

Andrew Davies is senior analyst for defence capability at ASPI and executive editor of The Strategist. Image courtesy of Department of Defence.

Mega fauna and the DCP

Mega fauna

Yesterday I appeared in front of the Joint Parliamentary Committee on Public Accounts and Audit. It’s not my usual haunt, but they were holding a public hearing into the Australian National Audit Office’s Audit Report No. 6 (2013-14) Capability Development Reform, which they correctly guessed I had an interest in.

Appearing in front of Parliamentary committees is always interesting. And it’s not a bad cure for cynicism caused by the sound bite focused 24/7 news cycle reporting of politics. Invariably, the committees are interested in understanding complex issues, and there’s rarely evidence of partisan point scoring.

Here’s what I told them. You’ll be able to read the exchange that followed in Hansard in a little while. (Complete with all of my misspeaking and false starts to sentences etc.—it’s not always edifying reading for the speaker!) Read more


I’d like to thank you for the opportunity to make a short opening statement. By way of background, I have been professionally involved in defence capability development for nearly two decades now. That has included providing scientific input into force development considerations, managing a sizeable major project, and the management and planning of a portfolio of projects within intelligence. Since leaving the Department, I have been a frequent contributor to the public debate on defence policy, planning and acquisitions.

But I should start with two caveats. First, it’s now almost eight years since I was an employee of the Department of Defence, and longer since I was involved in the central processes of capability development that are the subject of this Committee’s deliberations. While I attempt to stay as current as possible, there are bound to be points of detail I’m behind the curve on. Secondly, the Minister announced last week that I would be on his advisory panel for the development of the next defence white paper. That process has not begun yet, so anything I say today in no way reflects the views of the Minister or of the department of defence.

When the ANAO report into capability development was released late last year, my colleague Mark Thomson and I wrote a critique of it for our corporate blog. I’d like to table that work in two parts. For the public record, I’d add that I received some comments and a very helpful briefing from the Capability Development Group between the publication of those two parts. They felt that the first part reflected an insufficient appreciation of the efforts made to reform defence capability planning in recent years. The second part, I hope, reflects the discussion we had.

Let me say that the criticisms I’m about to make shouldn’t be taken as a reflection of either the professionalism or the hard work of the CDG. Rather, I think they have been grappling with some difficult issues, and making some decent headway towards a more rigorous and internally transparent process. I say ‘internally’ there quite deliberately. From the outside it’s often difficult to discern the details of the process or to infer the train of logic underpinning decisions from the observed outcomes.

Because of that, I think the most productive way for me to proceed today is to largely ignore the internal machinations of the CDG and to look instead at their inputs and outputs. I think there are substantive issues at both ends of the process.

We can safely conclude that something is badly askew in the capability development process based on an assessment of the current Defence Capability Plan and the Defence budget. Mark Thomson has described the latter as an ‘unsustainable mess’—a turn of phrase not used lightly—and I refer you to his budget briefs of the last few years where you’ll find a great deal of evidence to support that claim.

My role at ASPI is to look at force structure planning, and I take a strong interest in the DCP and its capability and budgetary implications. And there’s a huge problem on the horizon. Simply put—there’s nowhere near enough money in the Defence budget, even if promised increases in funding manage to fight their way through the competing government priorities.

By any reasonable calculation, the DCP is heavily oversubscribed. I can’t recall a time in which there were so many very large projects. The DCP doesn’t have just one elephant in the room, it has a herd of them. The future submarine, future frigate, F-35 joint strike fighter, armoured vehicles for the Army and maritime patrol aircraft projects total, by my estimation, over $100 billion for the acquisition phase. To put that in perspective, the current annual acquisition budget is around $5 billion.

In other words, the ‘mega fauna projects’ would consume two decades worth of the current funding on their own—leaving little or no room for all of the other projects that are necessary to keep the ADF effective. This can’t work.

When Mark Thomson and I did our own costings of force structures we got a very different answer to the costings—such as they were—in the 2009 defence white paper. I’m happy to talk some more about that.

As I see it, the capability development process is effective, more or less, at de-risking individual projects, but isn’t adequately managing the portfolio of inputs. I suspect that there are problems with the inputs to the process—another point we might usefully discuss. Simply put, the best process in the world can’t produce sensible answers from the wrong starting point.

Andrew Davies is senior analyst for defence capability at ASPI and executive editor of The Strategist. Image courtesy of Flickr user benh LIEU SONG.

Graph of the week: if you want it on time, keep it simple

Late last year, I gave some early impressions of the Australian National Audit Office’s Major Projects report. Today I’ll have a closer look at some of the data in the report, which will bring us back to another of ‘Gumley’s Laws’. This time the subject is the impact of developmental work on project delivery. (The previous Gumley’s law post was on over-programming in the Defence Capability Plan.)

First, the data. The figure below is taken from the ANAO report and is DMO’s breakdown of the delay in projects delivering Final Operating Capability to the ADF. The projects are split into three categories—military off-the-shelf (MOTS), ‘Australianised MOTS’ and ‘developmental’.

Source: ANAO Major Projects Report 2012–13

Read more

The figure is a stark illustration of the penalty paid for developmental work. None of the developmental projects have been delivered without a 40% overrun in schedule, and the average is 92%—almost double the planned time. Note that the F-35 Joint Strike Fighter is excluded by the DMO as ‘not applicable’ because it’s ‘should ultimately become MOTS when it enters production line delivery’. But it certainly doesn’t buck the trend—on original plans the RAAF were to have their first operation squadron in 2012, but will now have to wait until 2020.

The breakdown into three categories is very coarse, but previous data sets had more fidelity, which allow us to estimate the effect of project complexity on schedule. The figure below is drawn from earlier ANAO reports, when the DMO gave a project complexity figure on a scale from 1 (simplest off-the-shelf) to 10 (most complex integration and development). While the data set is fairly sparse, it suggests the expected delay in project delivery is about 6–10 months per degree of complexity compared to an OTS purchase. And that’s about right when we compare it to the delays in delivering projects like the Collins class submarines and the F-35.

Source: DMO data on major defence projects

Gumley’s laws include two observations pertinent to this. The first is that:

Off-the-shelf acquisition of products that are technically mature is the only reliable way of avoiding cost and schedule risk. (Gumley’s 1st law)

That rule (which should be self-evident) has a natural corollary that Gumley also identified:

 Deviation from MOTS increases risk, cost, and schedule and should only be pursued when there are no MOTS systems with acceptable performance. (Gumley’s corollary to Law 1)

I’d go further than that. If we combine Gumley’s first law with Augustine’s observation that ‘the last 10% of the performance sought generates one-third of the cost and two-thirds of the problems’, I think the corollary should take Augustine into account and read more like this:

Deviation from MOTS increases risk, cost, and schedule and should only be pursued when there are no MOTS systems with performance within 10–20% of that sought and when the extra performance is judged to be critical.

Gumley’s second law is a less obvious, but no less important observation:

 Combining two MOTS systems is not MOTS. (Gumley’s 2nd law)

It’s important because it’s much easier to conceive projects that involve taking two separate MOTS components and integrating them —even ones that are themselves mature and fully functional—than it is to deliver them. Some examples from Australia’s experience include the lengthy process of integration of the AGM‑142 missile onto the Royal Australian Air Force’s F‑111 fleet and the much delayed FFG frigate upgrade. In both case all of the subsystems were well-understood, but the interplay between them was anything but straightforward to manage. The FFGs re-entered service years later than expected and the F-111 managed to fire its first AGM-142 just over a year before the airframe was retired. (A quick mea culpa: my fingerprints will be found on that sorry saga.)

Perhaps the most egregious example was the failed attempt to modernise the Super Seasprite helicopter, which required the integration of flight and combat systems, weapons and sensors sourced from multiple suppliers. The net result was a $1.4 billion investment for no capability return at all.

Gumley’s second law also has a corollary:

 ‘Australianised MOTS’ is an oxymoron.

Or, to put it simply: ‘if you change it, it isn’t MOTS’. Again, the data bears this out. The ‘Australianised MOTS’ projects in the ANAO chart above have an FOC delivery variance of just under 50% above the planned figure—not as bad as the developmental projects, but a long way short of the MOTS projects, for which the average slippage is closer to 2%.

Andrew Davies is senior analyst for defence capability at ASPI and executive editor of The Strategist

We’ll build ships together

Crane 8 at the old Cockatoo Island dockyards

It’s becoming something of an annual tradition that I get asked along to the Australian Defence Magazine (ADM) conference and get to be the token economic rationalist on Day 2. At an industry-focused conference, that’s a bit like being the crazy uncle at the Christmas dinner table. Last year, for example, I was asked to talk about whether there’s a future for non-sovereign providers of defence material—and decided instead to argue that it’s actually the sovereign providers who are an endangered species.

In other ventures onto the stage at defence industry functions I’ve variously argued that we should take advantage of other countries (PDF) that prop up their defence industries with government funding by buying the below cost products that emerge, and (with partners in thought crime Henry Ergas and Mark Thomson) that we should allow the cold winds of competition (PDF) to blow through the Australian naval shipbuilding industry in the name of efficiency. Read more

So it makes a bit of change for me to take a different tack this year. At my talk earlier today for the 2014 ADM Congress, my starting point was that the future of shipbuilding in Australia is assured because of political buy-in, so we really should be talking about the best way to manage it. It’s not that I’ve had a roadside conversion, but I think that’s an accurate reading of the strong indications evinced by both sides of politics. Just yesterday, the Ministers of Defence and Finance announced a review into the troubled Air Warfare Destroyer program—of which more anon—and commented that the exercise:

… is a demonstration of the Government’s commitment to working collaboratively and constructively to ensure we realise both the national security benefits and the long term benefits of the AWD Program for the Australian shipbuilding industry.

So it looks like we’re going to have a naval shipbuilding industry for some time to come. The trend was already that way, but I think that the recent demise of the local car industry made it pretty much a done deal. Already the sharks are circling, and various industry players and state governments are vying for position, trying to get guarantees for work to keep their declining manufacturing base active.

But vested interests generally don’t make for good public policy, so the next step should be to work out how to protect the taxpayer’s interests—and the state of the AWD program is a strong indication that those interests need protecting.  Yesterday’s announcement was no surprise to those of us who have followed the increasingly fraught progress of the program; Mark Thomson previously identified potential problems with the project’s ‘alliance’ structure and the Australian National Audit Office and the DMO themselves provided enough insight into the state of play in last December’s review of major projects to make it clear that there was a big problem. Simply put, the project has been chewing through money faster than expected, and producing ships slower than expected, and there seems to be a real chance that it will run out of money before the third ship is delivered. (Another ANAO Audit report due next month should give a more detailed picture.)

The first step is to accurately diagnose the problem, which should be the main focus of the new review. The reasons likely to be cited will be many and varied, and will include attempts to blame the Spanish designer Navantia, and there’ll be much gnashing of teeth about the ‘stop-start’ nature of shipbuilding in Australia and the resulting penalties in getting shipyards back up to speed after a period of relative inactivity. Most of the proffered explanations will contain a grain of truth, and the picture that emerges might be a complex one.

But there are clearly some significant shipyard productivity problems at the heart of the problem, probably exacerbated by overly-optimistic estimates of how quickly the shipbuilders could get up to speed in the AWD build (despite over $200 million being spent on studies and risk retirement before contract signature). Again, that’s not surprising—with the government acting as sole customer and owner of the main shipbuilder, the sector hasn’t exactly been competition central. At least it’s almost certainly better than the old government owned naval shipyards at Cockatoo Island which gave rise to not one but two Royal Commissions in 1903 and 1921.

There are many tens of billions of dollars of naval shipbuilding projects in the wings, in the form of the two at-sea replenishment ships, twelve future submarines, twenty new patrol vessels and eight future frigates in the Defence capability Plan. In fact, there’s enough work in the forward plan for industry and state governments (South Australia’s foremost) to be pushing the idea of a ‘rolling production program’ in which we build ships and submarines in perpetuity.

So the stakes are very high in this review. It’s very hard to see how a government that is being very protective of its budget could enter into a commitment of that sort on the basis of the performance it has seen so far. Make no mistake—the worse the AWD story, the more risk-averse the government is likely to be. While there’s political buy-in to sustain a shipbuilding industry in principle, in practice, funding might be hard to come by if it looks like sheltering systemic underperformance.

In particular, the AWD program has probably just done the most aspirational vision for the future submarine a major disservice. Unless there’s a clear path is identified towards a higher performing shipbuilding sector, the less risky submarine options can only become more appealing. We’ll tease out the implications of that here on The Strategist in the next few weeks.

Andrew Davies is a senior analyst for defence capability at ASPI and executive editor of The Strategist. Image courtesy of Flickr user The Waterboy.