January 29, 2015
Getting Faster at Revolutionary
The bureaucratic and technological challenges in speeding up weapons development
By James Hasik
Pentagon procurement czar Frank Kendall told the House Armed Services Committee yesterday that his department, as the headline went, "Needs to Get Faster at Developing Revolutionary Weapon Systems”. He also noted that the Air Force, in conjunction with DARPA and the Navy, would soon begin serious work on a “sixth-generation” fighter (marketing-speak for the next new thing). As Defense News reported yesterday, "Air Force officials have said they expect to begin flying the next-generation jets by as early as the 2030s.” That is to say, if they start now, they may take as little as fifteen years, or about as long as they have with the Joint Strike Fighter. With those kinds of contradictory statements, it’s worth asking—has the Defense Department given up before starting? Has all this recent talk about faster innovation been just that, because no one wants to admit to some inexorable barrier standing in the way? Administrative problems are frequently discussed, but fundamental technological challenges may prove more ominous.
Another possibility is that the problems slowing up weapons development are more technological, deterministic, and fundamental. As Kristin Oakley described in this column last year, the ratio between RDT&E spending and procurement spending has been gradually rising for Defense since at least the mid-1960s. Some of that trend is a matter of accounting; the software content of weapons systems has been continuously rising, and writing software is counted as development, whether or not the software techniques break new ground. But there’s a nagging feeling, as one former Net Assessment staff officer put it to me, that “a dollar does not deliver as much advance as in the past”. In the January 2009 Review of Economic Studies, Benjamin Jones of Northwestern’s Kellogg School argued that "accumulating knowledge” is demanding progressively greater investment for each succeeding advance. (See “The Burden of Knowledge and the ‘Death of the Renaissance Man’: Is Innovation Getting Harder?” vol. 76, no. 1, pp. 283-317). In support, he cites several observations about innovators’ behavior from a large dataset on patents and their inventors: the age at first invention, specialization, and the need for teamwork have all been increasing. This suggests that more high-class labor has been required for new developments over time.
Meanwhile, the leadership is veritably anguished that adversaries may be catching up. Michael Dumont, the principal deputy assistant secretary of defense for special operations/low intensity conflict, spoke on just this topic last week at the National Defense Industrial Association's Special Operations/Low Intensity Conflict conference. (Yes, the PDASD for SO/LIC spoke at SO/LIC.) “Many of our adversaries,” he observed, “have acquired, developed and even stolen technologies that have put them on somewhat equal footing with the West in a range of areas.” He specifically cited computer hacking. He then said something about “the austere budget environment that we currently face,” which makes one wonder how the hackers manage on less than a half trillion dollars a year. In all seriousness, it does beg the question—how can our people be moving so slowly in some fields, when our adversaries are moving so fast in others, and sometimes in the ones that seem to matter so much?
Whether in this case of cyber-conflict or more broadly, either the administrative or the technological explanation could work. Perhaps our people can’t hack hacking tools because they’re weighted down by bureaucracy. If so, and if adversaries are indeed catching up, then government officials should consider working all the harder at improvement initiatives—like that Better Buying Power. They should definitely disregard Harvey Sapolsky’s recent advice not to dwell on reform, for this would be dangerous defeatism. Instead, they would need to ask what the hackers are doing so well, and do some of that too. We might start with Lieutenant General David Deptula’s classic observation that “Al Qaeda doesn’t have a JCIDS process,” and then blow up the worst parts of the bureaucracy. Then again, perhaps our people can’t hack, and shouldn’t hack back, because the increasing complexity of information systems is creating vulnerabilities faster than they can be plugged. The problem would then be technological, and there would be only so much that could be done. In this field, as perhaps in others, adversaries would then long retain a second-mover advantage. By observing the technical fiascos generated by thousands of engineers pursuing unobtanium, they would spend far less money on development, take the sensible shortcuts, and put the savings into procuring the new kit in greater numbers.
The technological explanation is ominous, as it suggests that the powers-that-be are doomed to fight on a leveling battlefield. There’s a disheartening economic analog in Jones’ aforementioned paper. Perhaps following Tyler Cowen’s opinion, he concludes by speculating that the accumulating burden of knowledge means that economic growth is inexorably slowing. But looking around at the explosion of new products and services, I call perceptions of slowing economic growth a measurement problem: the awesomeness of each “insanely great” new product from Apple, for example, never gets meaningfully captured in GDP figures. This iPhone Argument frequently gets easily taken too far in military matters, but it bears mention again here. As insanely great as all this stuff is, there’s a lot of investment involved. It’s often noted that Apple may not spend a large percentage of its revenues every year on new product development, but that’s only because its revenues are vast. The company's $6 billion in annual spending is only somewhat less than the $8.5 billion that the US Army spends on all its RDT&E. That’s a lot of money for dreaming up new gadgets, and there’s no live-fire testing to pay for.
Not just Apple, of course, but computing companies large and small keep stuffing more functionality into gadgets of roughly the same size. Thus, the technology of everyday life seems to have been changing more rapidly than the iconic technologies—tanks, ships, and airplanes—of military force. Perhaps this is why the Army keeps updating its Abrams and Bradleys and Paladins with new electronics, but can’t make a strong case for replacing those vehicles with that wholly next new thing. Different areas of technological endeavor show differential speeds of advance, but the post-Cold War military budget is moving sideways. If the entire Army is to have only somewhat more money to spend developing new weapons as Apple spends on new phones, then what’s the chance that there’s some huge new thing about to be discovered? Little—unless defense ministries match the rhetoric of leveraging commercial technologies with actual practice. Propelling all those advances in computing and communications has been the recombination of technologies, the plugging-and-playing of existing or recently-developed hardware and software together by standardized interfaces and protocols. This is what has been producing that uncomfortable iPhone or iRobot Feeling—why can’t we get something like that here, if we’ve got it back home? Because we still keep getting it back home, whatever Cowen and Jones might predict.
So I offer a counter-thesis: the rate of military innovation may not slow, and may even pick up, but only with a sharp change. The trend that must simply end is the pursuit of increasingly grands projets with investments in ‘big science’ that lead to great leaps forward. Perhaps there’s run room left in software, whatever Moore’s Law may do, as the tools of software development catch up for a few decades. We may then not be facing any great stagnation—military or economic—for quite some time. But getting this really would mean no more Joint Strike Fighters. Getting this right might mean just saying no to this FA-XX or whatever it would be, and instead levering the power of networking to distribute lethality further. That would be not just an economic imperative, but a military one.
James Hasík is a senior fellow at the Brent Scowcroft Center on International Security.